Journal of Anxiety Disorders 28 (2014) 241–245

Contents lists available at ScienceDirect

Journal of Anxiety Disorders

The impact of hypervigilance: Evidence for a forward feedback loop夽

Matthew Kimble a,∗ , Mariam Boxwala a , Whitney Bean a , Kristin Maletsky a ,
Jessica Halper a , Kaleigh Spollen a , Kevin Fleming b
a
Middlebury College, Department of Psychology, Middlebury, VT, United States
b
Norwich University, Department of Psychology, Northfield, VT, United States

a r t i c l e i n f o a b s t r a c t

Article history: A number of prominent theories suggest that hypervigilance and attentional bias play a central role in
Received 9 July 2013 anxiety disorders and PTSD. It is argued that hypervigilance may focus attention on potential threats and
Received in revised form 4 December 2013 precipitate or maintain a forward feedback loop in which anxiety is increased. While there is considerable
Accepted 16 December 2013
data to suggest that attentional bias exists, there is little evidence to suggest that it plays this proposed but
critical role. This study investigated how manipulating hypervigilance would impact the forward feedback
Keywords:
loop via self-reported anxiety, visual scanning, and pupil size. Seventy-one participants were assigned to
Anxiety
either a hypervigilant, pleasant, or control condition while looking at a series of neutral pictures. Those
PTSD
Hypervigilance in the hypervigilant condition had significantly more fixations than those in the other two groups. These
Attentional bias fixations were more spread out and covered a greater percentage of the ambiguous scene. Pupil size was
Eye tracking also significantly larger in the hypervigilant condition relative to the control condition. Thus the study
provided support for the role of hypervigilance in increasing visual scanning and arousal even to neutral
stimuli and even when there is no change in self-reported anxiety. Implications for the role this may play
in perpetuating a forward feedback loop are discussed.
© 2014 Published by Elsevier Ltd.

1. Introduction all of which would further increase anxiety. Empirical evidence

for attentional bias across a range of anxiety disorders is quite
Past and present theories have suggested that attentional bias prevalent (Bögels & Mansell, 2004; Horley, Williams, Gonsalves,
toward threat plays a central role in anxiety disorders, including & Gordon, 2004; Schofield, Johnson, Inhoff, & Coles, 2012; Weeks,
posttraumatic stress disorder (Brewin, Dalgleish, & Joseph 1996; Howell, & Golden, 2013). Referring specifically to PTSD, Chemtob
Chemtob, Roitblat, Hamada, Carlson, & Twentyman, 1988; Cisler et al. (1988) argued for a threat detection network that was eas-
& Koster, 2010; Dalgleish, Moradi, Taghavi, Neshat-Doost, & Yule, ily potentiated and could initiate a positive feedback loop. The
2001; Ehlers & Clark, 2000; Foa & Kozak, 1986; Foa, Steketee, error in interpretation then provides subjective evidence for threat,
& Rothbaum, 1989; Litz & Keane, 1989). In many of these mod- thereby increasing threat-related arousal. It is argued that this
els, attentional bias and hypervigilance are thought to play a increased arousal facilitates greater attention toward threat and
critical role in the maintenance, and perhaps etiology, of the dis- decreases cognitions or behaviors that may inhibit anxiety. Rec-
order. Dalgleish et al. (2001) for example, argued that “anxiety ognizing the critical role of hypervigilance, Conoscenti, Vine, Papa,
leads to increased hypervigilance for threat, a greater level of and Litz (2009) updated the body of work regarding hypervigilance
threat detection leads to increased anxiety which, in turn, leads in PTSD and considered the symptom a gateway to posttraumatic
to increased hypervigilance, and so on in a vicious circle” (p. 541). disturbance.
Beck, Emery, and Greenberg (2005) later argued that anxiety dis- Supporting this theory are both long standing clinical evidence
orders were associated with a hypervigilance that led them to for hypervigilance (Kardiner & Spiegel, 1947) as well as ample
misinterpret ambiguous situations and exaggerate minor threats, empirical support for the existence of attentional biases in PTSD.
In visual search tasks (Pineles, Shipherd, Welsh, & Yovel, 2007),
dot probe tasks (Bryant & Harvey, 1997; Dalgleish et al., 2001),
and modified Stroop tasks (Beck, Freeman, Shipherd, Hamblen,
夽 This publication was made possible by an AREA Grant (R15 MH081276) to & Lacker, 2001; McNally, Kaspi, Riemann, & Zeitlen, 1990), there
Matthew Kimble from the National Institute of Mental Health. Thank you to Andrew has been evidence for both facilitation (i.e., increased detection of)
Hyatt for technical assistance on the project.
∗ Corresponding author at: Psychology Department, Middlebury College, Middle- as well as an interference (i.e., poor disengagement from) threat-
bury, VT 05753, United States. Tel.: +1 802 443 5402; fax: +1 802 443 2072. related stimuli (for mixed evidence or exceptions see Kimble, Frueh,
E-mail address: mkimble@[Link] (M. Kimble). & Marks, 2009; Pineles et al., 2007; Pollak & Tolley-Schell, 2003).

0887-6185/$ – see front matter © 2014 Published by Elsevier Ltd.

[Link]
242 M. Kimble et al. / Journal of Anxiety Disorders 28 (2014) 241–245

Evidence for facilitation and poor disengagement in PTSD has

been bolstered by recent work using eye tracking technology
(Beevers et al., 2011; Bryant, Harvey, Gordon, & Berry, 1995;
Felmingham, Rennie, Manor, & Bryant, 2011; Kimble, Fleming,
Brandy, Kim, & Zambetti, 2010). Eye tracking techniques can assess
patterns in eye fixations, fixation durations, and eye movement. In
addition, the technology provides continuous, non-invasive indices
of attention to visual stimuli. It has a distinct advantage over dot
probe, Stroop, visual search, and spatial cueing tasks in that it can
directly assess visual attention without the difficulties of inter-
pretation associated with reaction time. This early work in PTSD
has shown evidence for increased detection for threat with lit-
tle avoidance of threatening stimuli after detection (Bryant et al.,
1995; Felmingham et al., 2011; Kimble et al., 2010). In a recent
study, however, Beevers et al. (2011) found that predeployment
Fig. 1. Example visual stimulus with grid overlay.
avoidance of angry faces predicted the development of PTSD when
exposed to war zone stressors. The authors argued that avoidance
of fear-related information prior to deployment may indicate a ten- respond to (i.e., animals vs. faces vs. bridges), we thought it too
dency to avoid thinking about trauma and may later interfere with difficult to produce stimuli that would allow one to conclude a for-
the extinction and reappraisal processes necessary for recovery ward feedback loop in anxiety disorders more broadly. However,
(Beevers et al., 2011). Eye gaze was not assessed post-deployment we feel this study can reasonably serve as a case in point. Specif-
however, and thus it is unclear whether the subsequent PTSD was ically, a demonstration of a hypervigilant condition that produces
associated with the facilitation and poor disengagement found in greater visual scanning and autonomic arousal to stimuli that are
the previous studies. most relevant to those with PTSD would certainly have implica-
While there has been a preponderance of evidence for facilita- tions for understanding hypervigilance in other anxiety disorders.
tion and poor disengagement in PTSD, no studies have investigated For example, it would be reasonable to conclude that a hypervigi-
whether these attentional patterns perpetuate a forward feedback lant mindset that produces increased anxiety and visual scanning
loop that results in increased anxiety, increased search for threat, to stimuli related to PTSD would do the same for social phobia.
and increased autonomic arousal. The lack of work in this area
is partly due to the difficulty of assuring in clinical samples that 2. Methods and materials
any changes in anxiety or arousal are actually due to changes in
hypervigilance and not some other confounding factor such as 2.1. Participants
trauma type, medication status, or symptom severity. The difficulty
is further complicated by ceiling effects in which increasing hyper- The participants were 71 students from an undergraduate lib-
vigilance may be difficult in an already vigilant, clinical sample. eral arts school in the northeast. One student was removed due
However, providing such data in any sample would produce criti- to poor quality eye tracking data. Participants were recruited from
cal evidence in support of the forward feedback loop proposed in introductory psychology courses and were given research credit
PTSD theory. For example, if hypervigilance or attentional bias for for completing the experiment. The average age of participants
threat does not produce further anxiety or behavioral change then was 18.4 years; there were 50 females and 21 males; 75% of par-
a forward feedback loop would, in theory, stop at that point. ticipants were white, 6% were African American, 11% were Asian
The goal of this study was to see whether manipulating hyper- American, 1% were American Indian, and 7% reported they were
vigilance would affect a forward feedback loop as measured by biracial.
self-reports of anxiety, impact on visual attention, and influence on
autonomic arousal. In order to control for confounding variables 2.2. Procedure
and ceiling effects, non-clinical participants were used and were
randomly assigned to hypervigilant, pleasant, and control condi- Participants were asked to sign an informed consent, and sub-
tions. Hypervigilance was manipulated through an instruction set sequently administered the Spielberger State Anxiety Inventory
in which participants were told that they needed to find the threat (Spielberger, Gorsuch, Lushene, Vagg, & Jacobs, 1983), the Post-
in a computer presented ambiguous scene in order to avoid an traumatic Stress Scale (Foa, Riggs, Dancu, & Rothbaum, 1993), along
aversive consequence. The impact of this instruction set on anx- with a brief demographic survey. Upon completion of the instru-
iety, visual scanning, and pupil dilation was assessed. Consistent ments, all participants underwent an eye tracking task in which
with theories that suggest a forward feedback loop in PTSD and they were asked to look at 30 pictures on a computer screen for 10
anxiety disorders more generally (Beck et al., 2005; Chemtob et al., seconds each. The pictures were presented on a Dell 22 monitor
1988; Dalgleish et al., 2001; Litz & Keane, 1989), we predicted that and took up the entire screen. Each participant s head was stabilized
the hypervigilant manipulation would result in larger pupil sizes with a chin rest and their eyes were 57 cm away from the monitor.
(i.e., greater autonomic arousal), a greater number of fixations (i.e., The pictures were neutral in valence and included an assortment
increased scanning), and higher self-reported anxiety. of images such as streets scenes, parking lots, ballparks, forests,
In this study, we constrained the stimuli to be most relevant to stores, and classrooms (see Fig. 1 for an example stimulus). All pic-
those with trauma and PTSD. Most pictures were neutral scenes tures were purposefully neutral but complex in order to simulate
that those with PTSD typically describe as potentially threaten- real life situations in which those with PTSD often describe as
ing such as railway platforms, busy rooms, empty streets, parking threatening even in the absence of overt threat.
garages, and so forth. The stimuli were purposefully neutral (with- Unknown to the participants, they were randomly assigned to
out any obvious threat) and ambiguous given that those with PTSD one of three conditions. Participants in the “hypervigilant condi-
are not necessarily seeing real threats in the environment, but tion” (N = 25) were instructed to search each picture for threatening
rather perceiving threats that are not there. In addition, given the targets. They were told that if they did not find all targets, a loud
breadth of stimuli that those with anxiety disorders generally might white noise burst (90 dB) would be played. Participants in the
 M. Kimble et al. / Journal of Anxiety Disorders 28 (2014) 241–245 243

“pleasant condition” (N = 24) were instructed to search each pic- Table 1

Self report & physiological measures.
ture for pleasant targets and were also told that a loud noise would
be played if they did not find all of them. Participants in the “con- Control condition Pleasant Hypervigilant
trol condition” (N = 21) were instructed to look at each image for condition condition
ten seconds and ignore the loud white noise. Female N = 16 N = 16 N = 18
The “hypervigilant” condition was designed to emulate the cog-
Age M = 18.38 M = 18.57 M = 18.33
nitive processes that may underlie hypervigilance, i.e., look for SD = 1.02 SD = .0.66 SD = 0.70
threat (in this case the threatening targets) in order to avoid an
PSS M = 7.05 M = 5.48 M = 4.08
aversive consequence (in this case, the loud noise burst). This would SD = 9.17 SD = 7.77 SD = 5.10
be an experimental analog to an individual with PTSD who might
STAI before M = 38.81 M = 35.04 M = 37.44
look out for danger (i.e., look over their shoulder) in order to avoid
SD = 9.87 SD = 10.35 SD = 9.6
a negative consequence (i.e., getting attacked from behind). The
pleasant condition was designed to control for valence effects (it STAI aftera M = 42.19 M = 38.92 M = 41.36
SD = 12.52 SD = 10.00 SD = 8.55
was identical except participants were told to find the pleasant
stimulus), and the control condition was designed to cover possi- Fixationsb M = 28.01 M = 30.06 M = 32.73
SD = 3.96 SD = 4.11 SD = 3.95
ble effects due to performance anxiety, i.e., hearing the loud noise
suggesting that they were failing the task. % Grids hitb M = 18.90 M = 20.68 M = 22.56
SD = 3.24 SD = 3.36 SD = 3.23

Pupil sizec M = 1173.6 M = 1296.5 M = 1447.11

2.2.1. Task SD = 235.98 SD = 342.53 SD = 346.31
In fact, there were no actual targets in the pictures, nor were a
Anxiety scores increased from before to after in all three conditions.
the noise bursts tied to the participants’ performance. All partic- b
Hypervigilant condition was significantly different from both control and pleas-
ipants were presented all 30 pictures in exactly the same order. ant conditions.
Noise bursts were presented in a fixed random order. In all, 30 noise c
Hypervigilant condition was significantly different from control condition.
bursts were presented. Four trials had two noise bursts presented
during the ten seconds, 14 trials had one, and 12 had none. This pat-
[F(2,67) = .54, p > .05], pretest anxiety [F(2,70) = .86, p > .05], or PTSD
tern of presentation was chosen in order to give the participant the
scores [F(2,68) = .86, p > .05]. See Table 1.
sense that the task was possible, but at times they failed to find the
“targets.” For any given slide in which a burst was presented, the
noise bursts were presented at fixed times. All noise bursts occurred 3.2. Self report measures
between 2 and 9 seconds after picture onset, were presented at
90 dB, were 1 second long, and were presented through two Altec- A 3 (Group) × 2 (Time) repeated measures ANOVA indicated a
Lansing speakers (model BXR1120) placed 24 inches behind the significant main effect of “Time” for the STAI [F(1,67) = 11.01, p < .05]
participant. Therefore, picture presentations and noise bursts were with all three conditions increasing in STAI scores in the post-
identical for all participants; the conditions only differed in their task relative to the pretask. There was no significant Group × Time
instruction sets. interaction however suggesting the groups did not differ in their
self-reported anxiety as a function of group [F(2,67) = .028, p > .05].

2.2.2. Eye tracking 3.3. Behavioral measures

Eye fixations and pupil diameter were measured using an Eye-
link 2000 desktop, mirrorless eye tracker (SR Research: Ottawa, 3.3.1. Fixations
Canada). Saccades and fixations were measured using corneal A one-way ANOVA was used to analyze the number of fixations
reflection and pupil detection algorithm standard to the Eyelink and all physiological variables. A significant difference was found
2000 system. Data was sampled from the left eye at 2000 Hz. between groups in the average number of fixations to the pictures
Data Viewer® software was used to overlay fixation and pupil [F(2,69) = 8.06, p < .05]. LSD post hoc analyses indicated that the
size over the stimuli for later offline analyses. The eye tracking hypervigilant condition produced significantly more fixations than
procedure began with a calibration/validation sequence in which did the pleasant condition (p < .05) or the control condition (p < .05).
participants followed a dot across nine locations on the screen. The control condition did not differ significantly from the pleasant
The eye tracker was calibrated on a per subject basis at the begin- condition (p > .05).
ning of each experiment and checked for accuracy between each
trial. Therefore, the interstimulus interval varied and was usually 3.3.2. Visual scanning
between 2 and 5 seconds. Dependent measures included the num- A significant difference was found between groups in percent of
ber of fixations during the 10 second trial, pupil size (measured sectors hit [F(2,69) = 7.15, p < .05]. LSD post hoc analyses indicated
in pixels), and visual scanning (as measured by the percentage of that the hypervigilant condition had significantly more sectors hit
sectors hit out of 88 possible sectors: see Fig. 1). The entire eye than did the pleasant condition (p < .05) or the control condition
tracking task typically took 7–8 minutes. When the task was com- (p < .05). The pleasant condition did not differ significantly from
pleted, participants filled out the STAI a second time and then were the control condition (p > .05).
debriefed.

3.3.3. Pupil size

3. Results A significant difference was found between the groups in
average pupil size [F(2,69) = 4.34, p < .05] LSD post hoc analyses
3.1. Demographic measures indicated that the hypervigilant condition produced significantly
larger pupil sizes than the control condition (p < .01). The average
A chi square test indicated that there were no differences pupil size for the pleasant condition was between the hypervigilant
between the groups on gender (x2 (2) = .38, p > .05). A one-way condition and the control condition but did not differ significantly
ANOVA revealed no significant differences between groups on age from either (p > .05).
244 M. Kimble et al. / Journal of Anxiety Disorders 28 (2014) 241–245

4. Discussion Increasing fixations and doing so over a broader area is likely to

increase the odds of finding a possible threat. Consistent with the
4.1. Self-reports of anxiety previous interpretation however, it is possible that increasing fix-
ations over a broader area may also decrease anxiety if no threat is
Contrary to what the forward feedback hypothesis would pre- found. In this sense, hypervigilance may be perceived by an individ-
dict, those in the hypervigilant condition did not report more ual as a behavior with few costs. If a threat is found, one is prepared
anxiety after completing the task. Rather, all participants experi- to act; if a threat is not present, one can relax.
enced an increase in anxiety. Presumably, the aversive noise bursts
caused an increase in anxiety regardless of condition. The increased 4.3. Implications and limitations
anxiety could not be attributed to performance anxiety as individ-
uals in the control group were told simply to look at the pictures and This study used a non-clinical, low in anxiety sample in order to
ignore the noise bursts. While there were no condition-related dif- look at the effects of increasing hypervigilance on self-reports and
ferences in self-reported anxiety, there were differences in pupil behavior. While this was done purposefully in order to minimize
size and visual scanning patterns. Such a pattern suggests that confounds and ceiling effects, one cannot be sure whether high
the hypervigilant-related instruction set produced autonomic and levels of hypervigilance play the role in a forward feedback loop in
behavioral changes even in the absence of changes in self-reports. those with PTSD that they played in this non clinical sample. How-
It is also possible that the nature of the stimuli may result in a ever, hypervigilant behavior in trauma survivors is likely to occur on
pattern in which we saw increased visual scanning but no increases a continuum that ranges from vigilance to hypervigilance and does
in self-reported anxiety. In particular, all of the pictures were neu- not suddenly “appear” in clinical samples with PTSD (Conoscenti
tral and did not have any overt threat. It is possible that actual et al., 2009). For example, Kimble and colleagues (2012) found high
threatening targets (or pictures with high negative valence) may levels of hypervigilance in civilian settings in all returning veterans,
have been needed to produce increases in self-reported anxiety. It even those without PTSD (Kimble, Fleming, & Bennion, 2013). While
seems likely that the inability to find a threat even when “hypervig- those veterans with PTSD had the highest hypervigilance scores
ilant” may have reassured the participants that a threat was really (and probably high enough to be functionally problematic), even
not present, and therefore result in them not reporting more anx- veterans without PTSD had high hypervigilance scores compared
iety. This would certainly be consistent with some clinical reports to survivors of other types of traumas. This suggest that hypervig-
indicating that individuals engage in hypervigilance in order to ilance is likely a variant of vigilance that differs in frequency and
feel safe. In situations in which a hypervigilant mindset is engaged severity, but not kind—suggesting that non-clinical analogs would
in and threats are not found, one might expect a reduction (or at have much to add.
least no increase) in reported anxiety. In this sense, hypervigilant In addition, one might argue that this analog study is simply a
behavior would be negatively reinforced through the alleviation demonstration of a manipulation with little implication for clinical
of anxiety. For example, a rape survivor who carefully enters her anxiety. The findings, however, understood within the context of
home at night and checks it for signs of illegal entry may signifi- anxiety and PTSD theory, suggest two very important points. First,
cantly reduce her anxiety upon assuring herself that the house is cognitions that mimic hypervigilance led to increased scanning
safe. In this study, the presence of only neutral (i.e., safe) visual and autonomic arousal. Second, both autonomic arousal and visual
stimuli may simply not have been sufficient to increase conscious scanning has been theoretically linked (Dalgleish et al., 2001; Litz &
anxiety. Keane, 1989) and in some cases has been demonstrated (see Prins,
Kaloupek, & Keane, 1995 for review of relevant empirical research)
4.2. Pupil size and scanning behavior to lead to further symptoms. Therefore, the increased scanning and
arousal seen in this sample has important clinical implications.
While not affecting self-reported anxiety, the hypervigilant Previous work would suggest that the increased scanning would
condition was associated with differences in pupil size. Pupil dila- lead to the finding of more threat and/or negative interpretations
tion is thought to be an index of autonomic, sympathetic arousal of ambiguous stimuli. This would, of course, perpetuate a forward
and is larger when looking at arousing stimuli (Bradley, Miccoli, feedback loop.
Escrig, & Lang, 2008). In PTSD, increased sympathetic arousal is Our analog for hypervigilance was highly experimental. We
thought to play a role in cueing both reexperiencing and avoid- tried to create an instruction set that mimicked a hypervigilance
ance symptoms. If this is the case, then a hypervigilant mindset, that might be similar to that experienced by an individual with
even in the absence of threatening stimuli, may increase auto- PTSD. In particular, we tried to generate a “mindset” in the hyper-
nomic arousal. However the hypervigilant condition did not differ vigilant condition that would make the participant alert for possible
from the pleasant condition, nor did the pleasant condition differ threat. We did so by setting up experimental conditions in which
from the control condition. This suggests that receiving feedback the participant was told to look for threat in an ambiguous visual
that one was failing at a visual search task for positive stimuli stimuli. In addition, the subject was informed that, if he or she
was not sufficient to significantly affect autonomic arousal (pleas- failed, there would be a negative outcome. This clearly is a mere
ant versus control), but failing to find a potentially threatening approximation of what a hypervigilant individual is likely to be
stimulus was sufficient in this respect (hypervigilance versus con- experiencing while scanning for threat in the real world. In partic-
trol). ular, the largest difference would be in a clinical patient’s belief that
The hypervigilant condition had its most robust impact on something catastrophic might occur if he or she do not identify the
visual scanning behavior. Those in the hypervigilant condition had threat before it is too late. This could not be mimicked in the lab.
significantly more fixations and looked over a broader area than However, the fact that our relatively mild hypervigilant condition
those in either of the other two conditions. These findings con- impacted autonomic arousal and visual scanning in a low anxiety
tribute to the literature in that most previous studies on attentional sample suggests that hypervigilance may have at least similar if not
bias demonstrate that those with PTSD look first or look longest larger effects in a clinical sample.
at threatening material. These finding suggest that hypervigilance While empirical validation would be critical, we feel it is rea-
is also associated with increased efforts to find threat. While this sonable to hypothesize that similar patterns may be present in
is entirely consistent with the clinical presentation of hypervigi- disorders such as social phobia, animal phobia, and panic disorder.
lance, there has been no previous empirical evidence to this effect. Hypervigilance for signs of rejection (i.e., social phobia), a feared
 M. Kimble et al. / Journal of Anxiety Disorders 28 (2014) 241–245 245

stimulus (i.e., snake phobia), or interoceptive abnormalities (i.e., Dalgleish, T., Moradi, A. R., Taghavi, M. R., Neshat-Doost, H. T., & Yule, W. (2001).
panic disorder), may also lead to increased autonomic arousal and An experimental investigation of hypervigilance for threat in children and
adolescents with post-traumatic stress disorder. Psychological Medicine, 31(3),
perceptual scanning that maintains a forward feedback loop. 541–547.
Ehlers, A., & Clark, D. M. (2000). A cognitive model of posttraumatic stress disorder.
4.4. Conclusions Behavioural Research Therapy, 38(4), 319–345.
Felmingham, K. L., Rennie, C., Manor, B., & Bryant, R. A. (2011). Eye tracking and
physiological reactivity to threatening stimuli in posttraumatic stress disorder.
In summary, this study provided mixed support for the for- Journal of Anxiety Disorders, 25(5), 668–673.
ward feedback hypothesis. Hypervigilance did not impact on Foa, E. B., & Kozak, M. J. (1986). Emotional processing of fear: exposure to corrective
information. Psychological Bulletin, 99(1), 20–35.
self-reported anxiety but did impact on visual scanning and arousal
Foa, E. B., Riggs, D. S., Dancu, C. V., & Rothbaum, B. O. (1993). Reliability and valid-
in a manner that could exacerbate other symptoms and result in ity of a brief instrument for assessing post-traumatic stress disorder. Journal of
further threat detection. Therefore, hypervigilance is a candidate Traumatic Stress, 6, 459–473.
Foa, E. B., Steketee, G., & Rothbaum, B. O. (1989). Behavioral/cognitive con-
to initiate or maintain a forward feedback loop through increased
ceptualizations of post-traumatic stress disorder. Behavior Therapy, 20,
scanning and autonomic arousal and theoretically could potenti- 155–176.
ate other difficulties even when the behavior does not result in Horley, K., Williams, L., Gonsalvez, C., & Gordon, E. (2004). Face to face: visual scan-
conscious increases in anxiety. path evidence for abnormal processing of facial expressions in social phobia.
Psychiatry Research, 127(1–2), 43–53.
Kardiner, A., & Spiegel, H. (1947). War stress and neurotic illness (2nd ed.). Oxford,
References England: P.B. Hoeber.
Kimble, M. O., Fleming, K., & Bennion, K. (2013). Contributors to hypervigi-
Beck, A. T., Emery, G., & Greenberg, R. L. (2005). Anxiety disorders and phobias: a lance in a military and civilian sample. Journal of Interpersonal Violence, 28(8),
cognitive perspective. New York, NY: Basic Books. 1672–1692.
Beck, J. G., Freeman, J. B., Shipherd, J. C., Hamblen, J. L., & Lacker, J. M. (2001). Speci- Kimble, M. O., Fleming, K., Brandy, C., Kim, J., & Zambetti, A. (2010). Eye tracking and
ficity of Stroop interference in patients with pain and PTSD. Journal of Abnormal visual attention to stimuli in veterans of the Iraq war. Journal of Anxiety Disorders,
Psychology, 110(4), 536–543. 24(3), 293–299.
Beevers, C. G., Marti, C. N., Lee, H., Stote, D. L., Ferrel, R. E., Hariri, A. R., et al. (2011). Kimble, M. O., Frueh, B. C., & Marks, L. (2009). Does the modified Stroop effect exist
Associations between serotonin transporter gene promoter region (5-HTTLPR) in PTSD? Evidence from dissertation abstracts and the peer reviewed literature.
polymorphism and gaze bias for emotional information. Journal of Abnormal Journal of Anxiety Disorders, 23(5), 650–655.
Psychology, 120(1), 187–197. Litz, B. T., & Keane, T. M. (1989). Information processing in anxiety disorders: appli-
Bögels, S. M., & Mansell, W. (2004). Attention processes in the maintenance and treat- cation to the understanding of post-traumatic stress disorder. Clinical Psychology
ment of social phobia: hypervigilance, avoidance and self-focused attention. Review, 9, 243–257.
Clinical Psychology Review, 24(7), 827–856. McNally, R. J., Kaspi, S. P., Riemann, B. C., & Zeitlin, S. B. (1990). Selective processing
Bradley, M. M., Miccoli, L., Escrig, M. A., & Lang, P. J. (2008). The pupil as a measure of of threat cues in posttraumatic stress disorder. Journal of Abnormal Psychology,
emotional arousal and autonomic activation. Psychophysiology, 45(4), 602–607. 99(4), 398–402.
Brewin, C. R., Dalgleigh, T., & Joseph, S. (1996). Dual representation theory of post- Pineles, S. L., Shipherd, J. C., Welch, L. P., & Yovel, I. (2007). The role of attentional
traumatic stress disorder. Psychological Review, 103(4), 670–686. biases in PTSD: is it interference or facilitation? Behaviour Research and Therapy,
Bryant, R. A., & Harvey, A. G. (1997). Attentional bias in posttraumatic stress disorder. 45, 1903–1913.
Journal of Traumatic Stress, 10(4), 635–644. Pollak, S. D., & Tolley-Schell, S. A. (2003). Selective attention to facial emo-
Bryant, R. A., Harvey, A. G., Gordon, E., & Barry, R. J. (1995). Eye movement and tion in physically abused children. Journal of Abnormal Psychology, 112(3),
electrodermal responses to threat stimuli in post-traumatic stress disorder. 323–338.
International Journal of Psychophysiology, 20(3), 209–213. Prins, A., Kaloupek, D. G., & Keane, T. M. (1995). Psychophysiological evidence for
Chemtob, C., Roitblat, H. L., Hamada, R. S., Carlson, G. J., & Twentyman, C. T. (1988). autonomic arousal and startle in traumatized adult populations. Philadelphia, PA:
A cognitive action theory of post-traumatic stress disorder. Journal of Anxiety Lippincott Williams & Wilkins Publishers.
Disorders, 2(3), 253–275. Schofield, C. A., Johnson, A. L., Inhoff, A. W., & Coles, M. E. (2012). Social anxiety
Cisler, J. M., & Koster, E. H. W. (2010). Mechanisms of attentional biases towards and difficulty disengaging threat: evidence from eye-tracking. Cognition and
threat in anxiety disorders: an integrative review. Clinical Psychology Review, Emotion, 26(2), 300–311. [Link]
30(2), 203–216. Spielberger, C. D., Gorsuch, R. C., Lushene, R. E., Vagg, P. R., & Jacobs, G. A. (1983).
Conoscenti, L. M., Vine, V., Papa, A., & Litz, B. T. (2009). Scanning for danger: read- Manual for the State-Trait Anxiety Inventory. Palo Alto: Consulting Psychologists
justment to the noncombat environment. In: S. M. Freeman, B. A. Moore, & A. Press.
Freeman (Eds.), Living and surviving in harm’s way: a psychological treatment Weeks, J. W., Howell, A. N., & Goldin, P. R. (2013). Gaze avoidance in social anxi-
handbook for pre- and post-deployment of military personnel (pp. 123–145). New ety disorder. Depression and Anxiety, 30(8), 749–756. [Link]
York, NY: Routledge/Taylor & Francis Group. da.22146