J Neural Transm (2009) 116:717–723

DOI 10.1007/s00702-008-0101-0

BASIC NEUROSCIENCES, GENETICS AND IMMUNOLOGY - ORIGINAL ARTICLE

Fear of negative evaluation and the hypervigilance-avoidance

hypothesis: an eye-tracking study
Matthias J. Wieser Æ Paul Pauli Æ Peter Weyers Æ
Georg W. Alpers Æ Andreas Mühlberger

Received: 8 July 2008 / Accepted: 20 July 2008 / Published online: 9 August 2008
Ó Springer-Verlag 2008

Abstract The hypervigilance-avoidance hypothesis disorders (e.g., Clark and Wells 1995). The vigilance-
assumes that anxious individuals initially attend to and avoidance hypothesis suggests that hypervigilance to the
subsequently avoid threatening stimuli. In this study pairs of feared stimulus is followed by attentional avoidance of it
emotional (angry or happy) and neutral facial expressions (e.g., Mogg et al. 1997). Several experimental paradigms
were presented to students of high or low fear of negative have been used to investigate the attentional processes in
evaluation (FNE) while their eye movements were recorded. social anxiety: the emotional Stroop task, the modified dot-
High FNE participants initially looked more often at emo- probe task, and visual search tasks (for a review, see
tional compared to neutral faces, indicating an attentional Bögels and Mansell 2004).
bias for emotional facial expressions. This effect was further Findings from the emotional Stroop tasks suggest that
modulated by the sex of the face, as high FNE clearly showed there is an initial vigilance bias for threatening stimuli in
a preference for happy female faces. Analysis of the time social anxiety (e.g., Spector et al. 2003). However, these
course of attention revealed that high FNE looked at the findings rely on threatening words, which may involve
emotional faces longer during the first second of stimulus semantic processing. Thus, effects may not solely be linked
exposure, whereas they avoided these faces in the consecu- to attention. Using ecologically more valid stimuli (i.e.,
tive time interval from 1 to 1.5 s. These results partially photographs of faces) in dot-probe tasks may more directly
support the hypervigilance-avoidance hypothesis and addi- assess attentional biases. Comparing reactions to words and
tionally indicate the relevance of happy faces for high FNE. facial expressions in a dot-probe task, social phobic par-
Further research should clarify the meaning of happy facial ticipants were found to show an attentional bias only for
expressions as well as the influence of the sex of the observed angry faces (Pishyar et al. 2004). Further studies confirmed
face in social anxiety. an early attentional bias in FNE persons toward angry
facial expressions (e.g., Mogg et al. 2004b). In contrast,
Keywords Fear of negative evaluation
there are also reports of an initial avoidance of threatening
Hypervigilance-avoidance
Eye-tracking
Social anxiety facial expressions in socially anxious students (Mansell
et al. 1999) as well as in clinically diagnosed social phobic
patients (Chen et al. 2002).
Introduction
Results from visual search tasks provide further insights
in attentional biases in social anxiety. Several studies using
Cognitive theories on anxiety assume that cognitive biases
facial expressions found evidence of enhanced attention
are related to the development and maintenance of anxiety
toward negative social cues (i.e., angry facial expressions)
in social anxiety (Perowne and Mansell 2002) and social
M. J. Wieser
P. Pauli
P. Weyers
G. W. Alpers
phobia (Gilboa Schechtman et al. 1999). However, there
A. Mühlberger (&) are also studies which failed to find evidence for such an
Department of Biological Psychology, Clinical Psychology
attentional bias in social anxiety (e.g., Esteves 1999).
and Psychotherapy, University of Würzburg, Marcusstr. 9-11,
97070 Würzburg, Germany Overall, results point towards the hypothesis that social
e-mail: muehlberger@[Link] anxiety may be associated with hypervigilance for

123
718 M. J. Wieser et al.

threatening social stimuli. However, it is noteworthy that Methods

these effects could also be explained by disengagement
processes (Koster et al. 2004) which may indicate that Participants
there is little impact of anxiety on initial detection of threat,
but on the disengagement from threat. Thirty female students of the University of Würzburg
Taken together, the described results point at vigilance completed the German version of the Brief Fear of Nega-
for potential threat in socially anxious participants, whereas tive Evaluation Scale (Leary 1983),1 of the Social Phobia
the data for the avoidance of these stimuli are weak. and Anxiety Inventory (SPAI, Turner et al. 1989; Fydrich
Because each of these paradigms assesses attention in an 2002), the German version of the State-Trait Anxiety
indirect way, further cognitive processes could have inventory (STAI, Spielberger et al. 1970; Laux et al. 1981),
influenced the results. Additionally, these tasks do not and a questionnaire on sociodemographic variables.
allow evaluating the exact time course of attention allo- Participants were divided by a mediansplit on their BFNE
cation. While they point towards the assumption of an scores in groups of high fear of negative evaluation (high
initial vigilance for threat in social anxiety, it is impossible FNE) and low fear of negative evaluation (low FNE) par-
with these paradigms to separate early from continuing ticipants. The BFNE is a self-report scale composed of 12
stages of attention. items describing fearful or worrying cognition. On a Likert
Probably the most direct method to assess selective Scale ranging from 1 (‘‘Not at all’’) to 5 (‘‘Extremely’’) the
attention and the time course of attentional deployment is respondents indicate the extent to which each item describes
to continuously record the exact position of eye gaze by them. Eight of the twelve items describe the presence of fear
eye tracking devices. The eye tracking methodology pro- or worrying (straightforwardly worded, or positively scored
vides an excellent research tool to investigate a variety of items), while the remaining four items describe the absence
processes in visual tasks related to attention (for a review, of fear or worrying (reversely scored items). This scale
see Findlay and Gilchrist 2003) and allows the discrimi- measures fear of negative evaluation, which is one of the
nation of early and late stages of attentional processing. hallmark criteria for the diagnosis of Social Phobia, and is
Few studies used eye tracking in order to assess the relevant to the study of human social behavior in general.
effects of general and social anxiety on attentional pro- Composed of questions nearly verbatim derived from the 30-
cesses during viewing of emotional pictures (Bradley et al. item Fear of Negative Evaluation (FNE) Scale (Watson and
2000; Mogg et al. 2000; Garner et al. 2006). In two studies, Friend 1969; Vormbrock and Neuser 1983) the 12-item
gaze behaviour was recorded while a conventional dot- BFNE Scale has the practical advantage of brevity, and has
probe task was performed (Bradley et al. 2000; Mogg et al. become a frequently used instrument in social anxiety
2000). In patients suffering from generalized anxiety dis- research (Rodebaugh et al. 2004). As the 12 items of the
order (GAD), but not in controls and depressed disorder English version of the BFNE, we took the 12 corresponding
patients, an enhanced probability to initially fixate threat- items from the German version of the FNE (Vormbrock and
ening compared to neutral faces was found (Mogg et al. Neuser 1983), as there is no published German version of the
2000). This bias in eye movements was also observed in BFNE itself. In past research on clinical and non-clinical
high trait anxious persons, but only in participants who populations the English version of the BFNE has demon-
performed gaze shifts at all (Bradley et al. 2000). Recently, strated good psychometric properties with an alpha
in two experiments it was found that only under social- coefficient of 0.88–0.96 (Leary 1983; Rodebaugh et al.
evaluative stress conditions (anticipation of a speech), high 2004). In our sample, the internal consistency was 0.93.
FNE participants showed a rapid initial orienting to emo- Furthermore, the high correlation between BFNE and SPAI
tional (both happy and angry) compared to neutral faces, scores in this study, r = 0.75, and a recently published study,
but showed reduced maintenance of attention on emotional r = 0.75, (Wieser et al. 2008) demonstrates the good
faces (Garner et al. 2006). external validity of the BFNE and therefore the practicability
The current study examines the attentional processes in and usefulness of the BFNE as a screening instrument for
FNE in response to emotional facial expressions (angry and social anxiety. One participant (high FNE) left the experi-
happy) by means of eye tracking in a passive viewing mental session early, so there were data of 29 participants
paradigm. We expected that high FNE participants would taken into analysis. High FNE participants were significantly
exhibit an early attentional bias for threatening facial younger (M = 20.7, SD = 1.8) compared to low FNE
expressions compared to low FNE participants. We further
1
followed up the time course of attention in order to As the low FNE group was significantly older than the high FNE
group, all the analyses of eye movement data were repeated using age
examine the hypervigilance-avoidance hypothesis. Alto-
as a covariate. The findings from analyses of covariance (ANCOVAs)
gether, this study aimed at a further elucidation of corroborated those from the ANOVAs by showing the same
attentional processes in social anxiety. significant effects and interactions.

123
Fear of negative evaluation and the hypervigilance-avoidance hypothesis 719

(M = 24.6, SD = 6.7), t(28) = 2.1, P = 0.045, and had pleasant’’ for the valence rating, and from 1 = ‘‘very
higher scores in the BFNE than low FNE participants, arousing’’ to 9 = ‘‘not arousing at all’’ for the arousal
t(28) = 2.1, P = 0.044, (M = 44.1, SD = 11.25, and ratings.
M = 34.7, SD = 10.9, respectively). No differences were During the experimental session, participants’ eye
found in measures of habitual and actual anxiety as measured movements were monitored continuously by a remote eye-
by the STAI state (low FNE: M = 37.7, SD = 5.6; high tracking recording system (iView X Hi-Speed, SensoMo-
FNE: M = 34.1, SD = 4.3) and STAI trait (low FNE: toric Instruments, Berlin, Germany). Eye movement
M = 40.2, SD = 12.4; high FNE: M = 42.2, SD = 11.3). measures were based on the corneal reflection caused by an
Furthermore, groups did not differ in their mean SPAI scores infrared light source (energy 3.5 mW/cm2 at a distance of
(low FNE: M = 102.0, SD = 26.4; high FNE: M = 108.6, 100 cm). Movements of the right pupil were recorded con-
SD = 31.1). tinuously at 240 Hz while subjects screened the displays.
All participants were without a history of any psychi- The eye-tracking system includes a chin and forehead rest to
atric and neurological disorder and reported normal or minimize head movements and to ensure a distance of 50 cm
corrected-to-normal vision. All of them gave their between the participant’s eyes and the centre of the display
informed consent and received course credits for their screen. Before each experiment, a standardized calibration
participation. The protocol for the study was approved by procedure for eye movements was performed by requesting
the ethics committee of the Deutsche Gesellschaft für participants to focus on 13 yellow crosshairs inserted in
Psychologie (German Association of Psychology). random order on the screen’s display space. Within an
experimental trial, the face pictures were presented only
Stimulus materials and apparatus when the subject focused exactly on the fixation cross,
ensuring that the initial position of the retina was controlled
Facial expressions of anger, happiness, and a correspond- during each trial across all participants. This was achieved by
ing neutral facial expression served as visual stimuli (for defining a circular area with a diameter of 1 cm around the
example, see Weyers et al. 2006). These facial expressions centre of the fixation cross. The presentation of the pictures
were created using the Poser 5 animation software (Curi- started only when the participants’ gaze matched the spatial
ous Labs, Santa Cruz, CA). This software combined with coordinates of this circle for at least 500 ms.
an extension by Spencer-Smith et al. (2001) allows creating
computer-generated facial expressions in an easy and Procedure
controlled way. The emotional facial expressions were
created from a prototype female and male face (for details After arriving at the laboratory, participants completed the
see Spencer-Smith et al. 2001). Each emotional expression informed consent form and the questionnaires described
was combined with three different styles (hairstyles and above. Then, they were comfortably seated in a chair in
hair colors) for the female and the male face, resulting in front of the monitor. Head movements were restrained by
18 stimuli (3 emotional expressions 9 3 types 9 2 sexes). the chin and forehead rest. Afterwards, participants were
The pictures were presented in pairs. Each pair consisted instructed to follow the instructions presented on the
of two pictures of the same person, with one picture depicting screen. After the calibration procedure and the practice
an emotional expression and the other a neutral expression. trials the experimental session started with the free viewing
The centers of the two pictures were separated by 8.9 cm. task. On each trial, first, a fixation cross appeared for at
Pictures were 13.6 cm high and 10 cm wide (7.7 and 5.7° of least 500 ms (see above) at the center of the screen, and
visual angle, respectively). The position of the target picture participants were asked to fixate it. Then, the pair of facial
(the emotional face) was balanced over the trials. Each stimuli was presented for 3 s and participants were free to
emotional expression was combined with the corresponding visually explore it. Afterwards, a blank screen was pre-
neutral one, resulting in 12 pairs. Pairs were presented four sented for a variable time interval between 500 and
times each in random order resulting in 48 trials. The pictures 2,000 ms before the next trial started with the presentation
were presented on a dark grey background on a 17 in. of the fixation cross. Finally, the participants were asked to
monitor with a resolution of 1,024 9 768 pixels. Partici- perform the rating procedure.
pants were instructed to fixate a fixation cross on the centre of
the screen until the two pictures appeared, and then to look at Data analysis
the pictures in any manner they chose.
After the tasks, the participants viewed each facial The purpose-written software package Begaze (SMI, Berlin)
expression once again and were asked to rate each picture on was used for analysis of eye movement data. The following
9-point Likert-scales according to arousal and valence. parameters of gaze behavior were calculated: the probability
Scales ranged from 1 = ‘‘very unpleasant’’ to 9 = ‘‘very of the first fixation on an emotional picture, and the latency

123
720 M. J. Wieser et al.

and duration of the first fixation on an emotional picture. The probability of fixation on emotional pictures (for angry as
probability of the first fixation on the emotional picture was well as for happy faces) than low FNE participants.
supposed to index initial orienting of attention. If eye However, the analysis revealed also a three-way interaction
movements appeared at least 100 ms after picture onset and Group 9 Emotion 9 Sex of the presented face,
were stable within a circular area of 0.5 cm diameter for at F(1,27) = 6.2, P = 0.020, g2p = 0.19. To clarify this
least 100 ms, this was classified as a fixation. First fixations interaction, separate follow-up ANOVAs for each group
were classified as being directed on a picture if they were were conducted. Only in high FNE participants the inter-
within the rectangular space of the pictures. The probability action Emotion 9 Sex was found to be significant,
of the first fixation on the emotional picture was expressed as F(1,13) = 17.1, p \ 0.001, g2p = 0.67, indicating that high
the relative frequency of fixation on this picture compared to FNE looked more often initially on female happy and
all fixations on a picture (i.e., fixations on the emotional angry male faces (see Fig. 1).
picture divided by fixations on the emotional and control
picture). Consequently, a probability above 50% indicates Attentional engagement over time
that the emotional picture was more likely the target of the
first fixation. Over the six 500 ms intervals of picture presentation, there
To follow-up the time course of the eye movements, the was a nearly significant interaction Group 9 Time,
3 s of picture presentation were subdivided into six 500 ms F(5,27) = 2.39, P = 0.08, g2p = 0.08 (see Fig. 2). This
intervals, and the proportion of viewing time directed on interaction was further analyzed by separate ANOVAs for
the emotional picture was computed (viewing time of the each time interval. For the first (0–500 ms) and the second
emotional picture divided by the total viewing time of the time interval (500–1,000 ms) the groups differed in their
emotional and the neutral picture). It was further assumed duration of fixations on the emotional pictures,
that gaze duration during the first two 500 ms intervals F(1,27) = 8.6, P = 0.01, g2p = 0.25, and F(1,27) = 5.4,
reflected early attentional engagement and hypervigilance, P = 0.02, g2p = 0.17, respectively. In both time intervals,
whereas the following four 500 ms intervals assessed late high FNE participants fixated emotional faces (angry and
attentional engagement and avoidance. Overall, a contin- happy) longer than did low FNE. This effect was reversed
uous measure of allocation of attention to emotional stimuli in the third time interval (1,000–1,500 ms), where low
was achieved. compared to high FNE tended to look at emotional facial
First fixation data were subjected to a repeated measures expressions for a longer amount of time, F(1,27) = 3.8,
analysis of variance (ANOVA) with group as between- P = 02.06, g2p = 0.13. For the following three time inter-
subject factor (2: high FNE vs. low FNE), and emotion (2: vals, no significant group differences were detected.
angry vs. happy), and sex (2: male vs. female face) as
within-subject factors. To analyze the time course of Ratings
viewing behavior, the proportion of viewing time was
analyzed by repeated measures analyses of variance In ratings of arousal and valence, no differences between
(ANOVA) with group as between-subject factor (2: high high and low FNE participants were detected. Concerning
FNE vs. low FNE), and emotion (2: angry vs. happy facial ratings of affective valence, angry faces were rated as to be
expression), sex (2: male vs. female face), and time (6: 0– more negative than either neutral or happy faces,
500 ms vs. 500–1,000 ms vs. 1,000–1,500 ms vs. 1,500– F(2,52) = 15.3, P \ 0.01, g2p = 0.32; contrast Fs
2,000 ms vs. 2,000–2,500 ms vs. 2,500–3,000 ms) as (1,26) [ 5.9, P \ 0.02, g2p [ 0.19, (angry: M = 2.3,
within-subject factors. Affective ratings were also analyzed SD = 0.8; neutral: M = 4.9, SD = 0.6; happy: M = 7.0,
by the repeated measures ANOVAs with group as between- SD = 1.1). Furthermore, female compared to male faces
subject factor (2: high FNE vs. low FNE), and emotion (3: were rated as more pleasant, F(1,27) = 8.6, P = 0.007,
angry vs. happy vs. neutral), and sex (2: male vs. female g2p = 0.24, (M = 4.9, SD = 0.5, and M = 4.6, SD = 0.7,
face) as within-subject factors. respectively).
For arousal ratings, there was a significant effect of
emotion found, F(2,54) = 11.0, P \ 0.001, g2p = 0.29.
Results Planned contrasts revealed that angry as well as happy
faces were rated as more arousing compared to neutral
Initial orienting (probability of first fixation) faces, F(1,27) = 3.4, P = 0.076, g2p = 0.11, and
F(1,27) = 41.4, P \ 0.001, g2p = 0.60, respectively,
For first fixation probability, there was a significant main (angry: M = 4.6, SD = 1.6; neutral: M = 3.9, SD = 1.4;
effect of group, F(1,27) = 10.3, P = 0.003, g2 = 0.29 (see happy: M = 5.5, SD = 1.4). Furthermore, female faces
‘‘Foot note 1’’), with high FNE participants showing higher were rated as more arousing than male faces,

123
Fear of negative evaluation and the hypervigilance-avoidance hypothesis 721

Fig. 1 Mean percentage of low FNE high FNE

initial fixation (+SEM) toward 70 70
emotional facial expressions
(angry and happy) in low FNE
female
male *
and high FNE groups (asterisks
60 60
indicate significant differences

emotional vs. neutral face

% initial fixations on
in post hoc comparisons)

50 50

40 40

30 30
0 0
angry happy angry happy
facial expression facial expression

70

low FNE
was found in response to female faces, only. These findings
high FNE provide evidence that happy facial expressions are of
% duration of fixations on emotional face

special interest in social anxiety, too. Regarding the time

*
60
* course of attentional deployment, a modest pattern of
* hypervigilant-avoidance related behavior was detected.
High compared to low FNE first attended more to emo-
tional faces regardless of emotion, whereas in the
50
consecutive time interval they directed their attention more
often to the neutral face. This is in line with the hyper-
vigilant-avoidance hypothesis (e.g., Mogg and Bradley
40 1998), but shows that emotional expressions per se are
initially attended and then probably avoided in social
anxiety. It has to be noted, however, that this avoidance
effect did not sustain and therefore might also be due to a
0
0-500 500-1000 1000-1500 1500-2000 2000-2500 2500-3000
compensatory reaction of the high FNE, which looked
time intervals (ms) longer at the emotional faces in the first second and then
turned to the neutral faces.
Fig. 2 Mean percentage of fixation duration (+SEM) on emotional Particularly the first fixation data, but also the analysis
facial expressions averaged over six consecutive time intervals for
low FNE and high FNE groups (asterisks indicate significant
of the time course of overt attention revealed that happy
differences between groups in post hoc comparisons) facial expressions are attention-grabbing for FNE persons.
Further research is needed to disentangle whether this
attention effect is because the happy expressions are
F(1,27) = 5.5, P = 0.027, g2p = 0.16, (M = 4.8, threatening for high FNE. This might be if these faces
SD = 1.1, and M = 4.6, SD = 1.1, respectively). indicate the potential start of a social interaction or might
be interpreted as a negative evaluation (e.g., mock or
‘‘Schadenfreude’’). Contrary, one might also speculate that
Discussion happy faces serve as security signals and therefore attract
attention. In both cases, our results point towards the
Using eye-tracking methodology, the attentional bias in suggestion that also happy facial expressions may be
response to emotional faces (angry and happy) was especially meaningful in social anxiety (e.g., Dimberg
investigated in high and low FNE participants. First fixa- 1997).
tion data revealed that high FNE participants show a Furthermore, this study revealed that the sex of the
hypervigilant pattern of attentional deployment in response presented face may modulate the attentional deployment of
to emotional faces (angry and happy), which is further high FNE to emotional expressions. Although the findings
modulated by the sex of the presented face. A clear are preliminary, as our sample consisted of women only,
attentional bias towards happy compared to angry faces these results show that the sex of faces should be

123
722 M. J. Wieser et al.

considered as an important variable in the research on of a speech), high FNE participants showed a faster initial
attentional biases in anxiety disorders. Thus, more studies orienting to emotional (both happy and angry) compared to
using male and female facial stimuli as well as male and neutral faces, and a reduced maintenance of attention on
female participants are needed to evaluate this issue, which emotional faces (Garner et al. 2006). This indicates that
interestingly has been neglected so far. situational aspects seem to modulate attentional deploy-
As a limitation of our study, it has to be noted that our ment and a mere sustained avoidance probably occurs only
results only apply for persons who have a higher fear of in situations of social stress.
negative evaluation, because the general social anxiety as Overall, results of this study point at an initial vigilance
measured by the SPAI did not differ significantly between for emotional facial expressions, which may be modulated
groups. Although fear of negative evaluation and general by the sex of the presented face. Interestingly, the bias
social anxiety are often highly associated, the attentional seems to be larger in response to happy female faces
biases to emotional facial expressions may be more pro- compared to angry female faces. Furthermore, hints for a
nounced in persons with high fear of negative evaluation as hypervigilance-avoidance pattern were found in the anal-
the main feature of anxiety. To generalize results on gen- ysis of the time course of attentional deployment. High
eral social anxiety, studies on persons with low and high FNE participants initially attended more to emotional than
levels of general social anxiety should be conducted. to neutral faces and consecutively this pattern reversed for
The use of eye tracking methods in research on emotion the second time interval. Further research should clarify the
processing and attentional biases seems to be a promising role of sex of the stimuli in attentional biases in social
tool to disentangle early and continuing components of anxiety. Additionally, including additional facial expres-
attentional biases or specific deficits in patients in response sions (e.g., fearful, sad) would be necessary to identify
to social stimuli. It has been shown in different psychiatric whether emotional faces per se capture attention in high
disorders, that measuring eye movements can disentangle FNE.
abnormal attentional processes in response to phobic
stimuli. For example, it has been shown, that social phobic Acknowledgment We thank Mathias Müller for technical assis-
tance. This work was supported by the grant MU2299-1/2 by the
patients avoid looking at facial features like the eye region Deutsche Forschungsgemeinschaft (DFG).
in sad and neutral faces and instead show a hypervigilant
scanning of relevant as well as irrelevant parts of the faces
(Horley et al. 2003). In autism disorder, it was found that References
autistic patients avoid looking at the eye region as well
(Corden et al. 2008). In spider phobic patients, it has been Bögels SM, Mansell W (2004) Attention processes in the mainte-
nance and treatment of social phobia: hypervigilance, avoidance
found that compared to healthy controls, they detect spiders and self-focused attention. Clin Psychol Rev 24:827–856
earlier (hypervigilance) and then avoid looking at them Bradley BP, Mogg K, Millar NH (2000) Covert and overt orienting
(avoidance) (Pflugshaupt et al. 2005). This was interpreted of attention to emotional faces in anxiety. Cogn Emot 14:789–
as evidence of a hypervigilant-avoidant attentional bias in 808
Chen YP, Ehlers A, Clark DM et al (2002) Patients with generalized
spider phobia. Likewise, our results support an early social phobia direct their attention away from faces. Behav Res
attentional bias and partially a later avoidance of emotional Ther 40:677–687
faces in high FNE. However, it remains unclear why this Clark DM, Wells A (1995) A cognitive model of social phobia. In:
avoidance behavior was found only in the time interval Heimberg RG et al (ed) Social phobia: diagnosis, assessment,
and treatment. Guilford Press, New York, pp 69–93
from 1,000–1,500 ms after stimulus onset. According to Corden B, Chilvers R, Skuse D (2008) Avoidance of emotionally
the hypervigilance-avoidance theory a longer lasting effect arousing stimuli predicts social-perceptual impairment in asper-
of avoidance would have been expected. Nevertheless, this ger’s syndrome. Neuropsychologia 46:137–147
period was also the critical one in dot-probe studies, in Dimberg U (1997) Social fear and expressive reactions to social
stimuli. Scand J Psychol 38:171–174
which these were interpreted as evidence for the avoidance Esteves F (1999) Attentional bias to emotional facial expressions. Eur
pattern (Mogg et al. 2004a, b). As our data show, this result Rev Appl Psychol 49:91–97
indeed may be an inherent methodological problem of the Findlay JM, Gilchrist ID (2003) Active vision: the psychology of
dot-probe paradigm, as only a snapshot of attention is looking and seeing. Oxford University Press, Oxford
Fydrich T (2002) Spai—Soziale Phobie und Angst Inventar. In:
measured by the reaction on the target. Thus, results are Brähler E, Schumacher J, Strauß B (eds) Diagnostische
always critically depending on the stimulus presentation Verfahren in der Psychotherapie. Hogrefe, Göttingen
time. However, in clinically diagnosed social phobic Garner M, Mogg K, Bradley BP (2006) Orienting and maintenance of
patients, no evidence of an attentional avoidance was found gaze to facial expressions in social anxiety. J Abnorm Psychol
115:760–770
at a presentation time of 1,250 ms (Mogg et al. 2004a, b). Gilboa Schechtman E, Foa EB, Amir N (1999) Attentional biases for
Furthermore, in two experiments it was found recently, that facial expressions in social phobia: the face-in-the-crowd
only under social-evaluative stress conditions (anticipation paradigm. Cogn Emot 13:305–318

123
Fear of negative evaluation and the hypervigilance-avoidance hypothesis 723

Horley K, Williams LM, Gonsalvez C et al (2003) Social phobics do Pishyar R, Harris LM, Menzies RG (2004) Attentional bias for words
not see eye to eye: a visual scanpath study of emotional and faces in social anxiety. Anxiety Stress Coping 17:23–36
expression processing. J Anxiety Disord 17:33–44 Rodebaugh TL, Woods CM, Thissen DM et al (2004) More
Koster EHW, Crombez G, Verschuere B et al (2004) Selective information from fewer questions: the factor structure and item
attention to threat in the dot probe paradigm: differentiating properties of the original and brief fear of negative evaluation
vigilance and difficulty to disengage. Behav Res Ther 42:1183– scale. Psychological Assessment 16
1192 Spector I, Pecknold JC, Libman E (2003) Selective attentional bias
Laux L, Glanzmann P, Schaffner P et al (1981) Das State-Trait- related to the noticeability aspect of anxiety symptoms in
Angstinventar. Weinheim: Beltz Test generalized social phobia. J Anx Disord 17:517–531
Leary MR (1983) A brief version of the fear of negative evaluation Spencer-Smith J, Wild H, Innes-Ker AH et al (2001) Making faces:
scale. Pers Soc Psychol Bull 9:371–375 creating three-dimensional parameterized models of facial
Mansell W, Clark DM, Ehlers A et al (1999) Social anxiety and expression. Behav Res Methods Instrum Comput 33:115–123
attention away from emotional faces. Cogn Emot 13:673–690 Spielberger CD, Gorsuch RL, Lushene RE (1970) State-trait anxiety
Mogg K, Bradley BP (1998) A cognitive-motivational analysis of inventory. Consulting Psychologists Press, Palo Alto
anxiety. Behav Res Ther 36:809–848 Turner SM, Beidel DC, Dancu CV et al (1989) An empirically
Mogg K, Bradley BP, De Bono J et al (1997) Time course of derived inventory to measure social fears and anxiety: the social
attentional bias for threat information in non-clinical anxiety. phobia and anxiety inventory (spai). Psychol Assess 1:35–40
Behav Res Ther 35:297–303 Vormbrock F, Neuser J (1983) Konstruktion zweier spezifischer trait-
Mogg K, Millar N, Bradley BP (2000) Biases in eye movements to Fragebogen zur Erfassung von Angst in sozialen Situationen
threatening facial expressions in generalized anxiety disorder (SANB und SVSS); construction of two specific trait-oriented
and depressive disorder. J Abnorm Psychol 109:695–704 questionnaires for the assessment of anxiety in social situations
Mogg K, Bradley BP, Miles F et al (2004a) Time course of attentional (SANB und SVSS). Diagnostica 29:165–182
bias for threat scenes: testing the vigilance-avoidance hypoth- Watson D, Friend R (1969) Measurement of social-evaluative
esis. Cogn Emot 18:689–700 anxiety. J Consult Clin Psychol 33:448–457
Mogg K, Philippot P, Bradley BP (2004b) Selective attention to angry Weyers P, Mühlberger A, Hefele C et al (2006) Electromyographic
faces in clinical social phobia. J Abnorm Psychol 113:160–165 responses to static and dynamic avatar emotional facial expres-
Perowne S, Mansell W (2002) Social anxiety, self-focused attention, sions. Psychophysiology 43:450–453
and the discrimination of negative, neutral and positive audience Wieser MJ, Pauli P, Alpers GW et al (2008) Is eye to eye contact
members by their non-verbal behaviours. Behav Cogn Psycho- really threatening and avoided in social anxiety?—An eye-
ther 30:11–23 tracking and psychophysiology study. J Anx Disord. doi:
Pflugshaupt T, Mosimann UP, von Wartburg R et al (2005) 10.1016/[Link].2008.04.004
Hypervigilance-avoidance pattern in spider phobia. J Anx Disord
19:105–116

123