Cognitive biases in anxiety

disorders and their effect

on cognitive-behavioral treatment
Michelle G. Craske, PhD
Deborah C. Pontillo, MA

Cognitive theorists hypothesize that cognitive biases are a major

component in the development and maintenance of anxiety
disorders. These include attentional biases toward threat-related
information, distorted judgments of risk, and selective memory
processing. The empirical evidence for these cognitive biases in
anxiety disorder populations is reviewed. Potential deleterious
effects of these biases on the process of cognitive-behavioral therapy
are also discussed, as are possible ways of overriding those effects
and maximizing treatment efficacy. (Bulletin of the Menninger
Clinic, 65[1], 58–77)

Cognitive processes have been incorporated extensively in conceptu-

alizations of the anxiety disorders over the past 2 decades. Cognitive
biases are defined as “any selective or nonveridical processing of
emotion-relevant information” (Mineka & Sutton, 1992, p. 65) and
tend to be emotion-, content-, and task-specific, in contrast to global
deficits in cognitive functioning. Thus the content of biases differs
from one anxiety disorder to the next. By generating further distress,
such biases are believed to contribute to the maintenance of excessive
anxiety and anxiety disorders. The source of these biases is not well
understood. Some speculate that they derive from core belief systems,
danger-laden in focus, that lead to distorted processing of informa-
tion about the world, self, and future through a filter of automatic
thoughts and images of danger (Beck, Emery, & Greenberg, 1985).
We first review the wealth of evidence for the presence of such
cognitive biases in anxiety disorders. In particular, we present the lit-
erature regarding attentional bias, memory bias, and judgment bias.

This article is based on a presentation at the 22nd Annual Menninger Winter

Psychiatry Conference held March 5–10, 2000, at Park City, Utah.
Dr. Craske is a professor of psychology in the Department of Psychology,
University of California, Los Angeles. Ms. Pontillo is a graduate student in the
Department of Psychology at the University of California, Los Angeles.
Correspondence may be sent to Dr. Craske at the Department of Psychology, UCLA,
405 Hilgard Avenue, Los Angeles, CA 90095-1563; e-mail: [email protected].
(Copyright © 2001 The Menninger Foundation)

58 Bulletin of the Menninger Clinic

 Cognitive biases in anxiety disorders

These biases are present, in varying degrees, in all the anxiety disor-
ders, although the content varies according to the anxiety disorder.

Attentional biases

Research has consistently demonstrated an attentional bias for

threat-related information in anxious populations. This has been op-
erationally defined as selective attention, perception, and encoding of
threat-relevant information (Williams, Watts, MacLeod, & Math-
ews, 1988). It is observed in anxiety disorder populations as well as
in nonclinical, trait-anxious populations (Broadbent & Broadbent,
1988). Moreover, these biases may be particularly salient when state
anxiety is elevated (Chen, Lewin, & Craske, 1996; MacLeod &
Mathews, 1988).
There are several mechanisms for attentional biases. One is strate-
gic or controlled attention in favor of threat-related stimuli. In con-
trast, nonanxious groups tend to divert attention away from threat-
relevant information (Eysenck, 1992). The controlled attention to
threat manifests clinically in overattention to negative stimuli that are
embedded in a larger positive context, such as the socially phobic
person who overattends to one audience member’s frown in an audi-
ence of mostly receptive faces, or the person with generalized anxiety
disorder who overattends to requests for minor corrections to a re-
port despite the generally positive tone of the feedback overall. This
is referred to as selectivity of attention, and was first demonstrated
experimentally using the dot probe detection task. In this paradigm,
neutral and positive valence words are presented and, at intervals un-
known to participants, one of the words is replaced with a target
“dot” requiring a response. Anxious individuals respond faster, in
comparison to nonanxious controls, when the target dot follows a
threat word of personal relevance (e.g., Mogg, Mathews, & Eysenck,
1992; Tata, Leibowitz, Prunty, Cameron, & Pickering, 1996).
Another mechanism for attentional bias is interference effects,
such as difficulty concentrating on task material when anxious
thoughts are present. Examples include the person with obsessive-
compulsive disorder who has difficulty attending to work demands
when experiencing intrusive thoughts, or the person with a phobia of
heights who cannot concentrate on the movie screen when seated in
the balcony row of a theater. The dichotic listening paradigm targets
this particular attention feature by requiring participants to verbally
repeat (“shadow”) nonthreatening information being played through
one ear of a pair of headphones, while threat and nonthreat words
are presented to the other ear. Anxious groups exhibit more disrup-

Vol. 65, No. 1 (Winter 2001)

Craske and Pontillo

tions of their shadowing task when threat words are presented to the
nonshadow ear (e.g., Foa & McNally, 1986; Mathews & MacLeod,
1986). Another measure of interference is the Stroop paradigm, in
which individuals are shown threat-related words and control words
printed in different ink colors. Individuals are asked to name the ink
color of the word. Significant delays in color naming are thought to
measure disproportionate allocation of attentional resources to the
word stimuli. Research has shown a substantial delay in naming col-
ors of threat-related words (modified Stroop task), and this delay is
pronounced in anxiety populations (e.g., Chen et al., 1996; Mathews
& MacLeod, 1985).
Attentional biases may also involve automatic or noneffortful,
nonstrategic processing biases toward threat-related information. A
consequence of this type of automaticity is experiencing elevated dis-
tress without verbalized awareness of threat. Thus the person with
panic disorder may experience an anxious reaction to changes in
heart rate without being fully cognizant of the change in heart rate;
that is, attention to heart rate occurred without conscious awareness.
This is demonstrated experimentally by presentation of stimuli in a
manner that prevents their conscious identification, such as at speeds
or decibel levels that are below thresholds for conscious detection
(MacLeod & Hagan, 1992; Mogg, Bradley, & Williams, 1995). Even
under these conditions, attentional biases are found toward person-
ally relevant stimuli in anxious populations.
As noted previously, attentional biases tend to be content specific
and, even more so, tied to personal relevance. For example, Logan
and Goetsch (1993) found that high anxious individuals demon-
strated greater Stoop interference for threatening information related
to current worries. Rieman and McNally (1995) showed that both
normal and anxious participants demonstrated a bias for positive
and negative stimuli that were concordant with current concerns.
Also, Mogg, Mathews, and Weinman (1989) found greatest interfer-
ence effects for threat words relevant to predominant worries of gen-
erally anxious individuals.

Memory biases

Whereas selective attentional bias and attentional interference for

threat-related stimulus have been demonstrated consistently, memory
biases are less clear in anxiety disorders. Bower’s (1981) hallmark
paper on mood and memory defined mood-dependent memory as fa-
cilitated retrieval when mood at retrieval is consistent with mood
during encoding. In addition, mood-congruent recall refers to facili-

60 Bulletin of the Menninger Clinic

 Cognitive biases in anxiety disorders

tated retrieval when the target material is mood congruent, regardless

of mood at encoding. Investigations of memory biases in depression
have consistently revealed a robust interaction between mood and
memory, especially for stimuli that are self-referential (for a review,
see MacLeod & Mathews, 1991). However, there have not been con-
sistent findings in the literature on memory biases in anxiety. For ex-
ample, Foa, McNally, and Murdock (1989) found no evidence for
mood-dependent memory, mood-congruent recall, or mood-congru-
ent encoding in anxious individuals. In terms of mood-congruent bi-
ases, Mogg, Mathews, and Weinman (1987) reported that anxious
patients recalled fewer anxiety words than did normal controls, as
did Watts, McKenna, Sharrock, and Trezise (1986). In a later study,
Mogg and Mathews (1990) did find a free recall mood-congruent
memory bias in individuals with high state anxiety compared with
controls. However, studies of recognition memory in clinically anx-
ious populations found an absence of a recognition memory bias fa-
voring threat (Mathews & MacLeod, 1985; Mogg, Gardiner,
Stavrou, & Golombok, 1992). Eysenck (1992) reviewed negative
findings from four out of five studies of recall and recognition biases
for threat versus nonthreat material in persons with generalized anxi-
ety disorder. Moreover, several studies with nonclinical, highly trait-
anxious and clinically anxious populations show no improved mem-
ory for anxiety-concordant information, and no anxiety-dependent
memory effects (e.g., Nugent & Mineka, 1994; Richards & Mill-
wood, 1989).
Conceivably, the failure to consistently find mood dependency and
mood congruency is due to inadequate testing conditions. Lang
(1985) suggested that it is similarity between physiological arousal
occurring at encoding and retrieval that mediates mood dependency
and mood congruency. Yet, in a test of mood congruency, McNally,
Foa, and Donnell (1989) found that even though panic disorder pa-
tients recalled more anxiety than nonanxiety words, this effect was
not significantly greater in panic disorder patients whose physiologi-
cal arousal had been increased. Eich (1995) suggests that for mood-
dependent memory effects to occur, mood at encoding and retrieval
must be similar in valence (positive vs. negative) as well as arousal
(high vs. low). In addition, he stresses the importance of a strong, sta-
ble mood, and internally generated retrieval processes in order to
evoke mood-dependent effects (i.e., free recall as opposed to recogni-
tion). In a recent study with students highly fearful of spiders or
snakes, we were able to demonstrate mood-dependent memory in re-
lation to states of fear and relaxation (Lang, Craske, Brown, &
Ghaneian, in press). Although replication is necessary, especially with

Vol. 65, No. 1 (Winter 2001) 61

Craske and Pontillo

clinical populations, our findings suggest that mood-dependent mem-

ory effects may exist in anxiety disorders under the specific condi-
tions outlined by Eich (1995).
The studies presented here refer to explicit memory, or the con-
scious retrieval of information, that is assessed through use of free re-
call, cued recall, or recognition tasks. Implicit memory, or priming, is
defined as performance change as a result of prior exposure without
reliance on conscious awareness. Implicit memory paradigms typi-
cally involve presentation of word stimuli ostensibly for a task ulte-
rior to learning, such as reading words aloud or making judgments of
words. Later, the previously seen words, along with new words, are
presented in a word task, such as word-stem completion or a percep-
tual identification task. The degree to which the previously seen
words affect performance on the word task is considered a measure
of implicit memory (Graf & Masson, 1993). In anxiety disorders, ev-
idence is more consistent for implicit memory than for explicit mem-
ory biases (e.g., MacLeod & McLaughlin, 1995; Mathews, Mogg,
May, & Eysenck, 1989). Enhanced implicit memory for threatening
information has been observed in nonclinical trait-anxious popula-
tions as well (Eysenck & Byrne, 1994; Richards & French, 1991,
1992). Nevertheless, some negative findings exist (e.g., Lundh,
Czyzykow, & Ost, 1997; Rapee, McCallum, Melville, Ravenscroft,
& Rodney, 1994).
Implicit memory biases are viewed as being concordant with atten-
tional biases, but involving different memory processes than explicit
memory biases. More specifically, distinctions are drawn between
processes required to activate a mental representation of a word and
processes required to further elaborate the meaning of a word (Graf
& Mandler, 1984). Activation is described as an automatic process in
which exposure to a word facilitates previously associated words.
Activation requires attentional resources and is sufficient for encod-
ing into implicit memory. However, it is not necessarily sufficient for
encoding into explicit memory. Elaboration is described as a strategic
process in which exposure to a word facilitates new associations with
additional words. Elaboration is considered necessary for encoding
into explicit memory.
Williams et al. (1988) propose that when anxious individuals are
exposed to threatening stimuli, additional fear-related associations
become automatically activated and further attentional resources are
allocated. Priming or implicit memory arises from these same auto-
matic processes. In light of the sparse evidence for explicit memory
biases, Williams et al. (1988) and others suggest that further process-
ing in the elaboration stage is avoided. In other words, anxious indi-

62 Bulletin of the Menninger Clinic

 Cognitive biases in anxiety disorders

viduals may be biased to attend to threat-related material but then to

avoid further encoding into memory (Mogg et al., 1987). Teleologi-
cal formulations reach a similar conclusion—the purpose of anxiety
is to identify potentially threatening stimuli rapidly rather than to re-
flect on past events (Mineka & Sutton, 1992).
However, the notion of enhanced activation (attentional and im-
plicit memory biases) and inhibited elaboration (explicit memory bi-
ases) is not well established. For example, Williams et al. (1988) did
not specify how processing during the automatic activation stage
could discriminate threat- from nonthreat-related information in
order to determine whether the later elaboration stage is to be
avoided. Furthermore, a test of attentional and memory processing in
which the elaboration hypothesis was directly tested failed to support
the model (Bradley, Mogg, & Williams, 1994), and directed forget-
ting paradigms have failed to show evidence for heightened avoid-
ance of encoding of threatening information (McNally, Metzger,
Lasko, Clancy, & Pitman, 1998; McNally, Otto, Yap, Pollack, &
Hornig, 1999). Finally, it does not take into account the possibility
that explicit memory biases do exist when measured under the right
conditions (Eich, 1995; Lang et al., 2000). Thus the relationship be-
tween attention and memory bias remains largely unresolved.

Judgment biases

Butler and Mathews (1983) hypothesized that heightened estimation

of personal risk culminates in the experience of anxiety, increased
salience, and thus increased availability of threat-related information.
Biases in the judgment of personal risk are measurable in terms of
statements of personal risk, judgments of negative associations, and
interpretations of ambiguous situations.
For example, Butler and Mathews (1983) presented negative (or
unpleasant) and positive (or pleasant) statements of events to highly
anxious individuals and control participants. All participants were
asked to subjectively rate the statements on a 0–8 scale of future
probability, where 0 = “not at all likely” and 8 = “extremely
likely.” Anxious individuals significantly overestimated subjective
personal risk for negative events when compared with controls, es-
pecially if those events were self-referent. The two groups did not
differ, however, on estimates of future positive events. Similarly, so-
cially anxious individuals judge negative social events to be more
likely and positive social events to be less likely than do controls
(e.g., Foa, Franklin, Perry, & Herbert, 1996), and acrophobic indi-
viduals overestimate the danger inherent in heights relative to con-

Vol. 65, No. 1 (Winter 2001) 63

Craske and Pontillo

trols (Menzies & Clark, 1995). A related concept is overestimation

of the fear response that characterizes phobic individuals (Taylor &
Rachman, 1994), which perhaps derives from an overprediction of
danger.
Biased perceived risk extends to judgments of negative associa-
tions. For example, Chan and Lovibond (1996) presented a series of
trials in which one stimulus was reliably paired with shock and the
other stimulus was never paired with shock. Participants rated their
expectancy for shock delivery. High-anxious college students, in con-
trast to controls, had higher expectancies for future shock trials for
stimulus configurations that were never paired with shock. This was
consistent with the authors’ previous research, which indicated con-
sistent overestimations of risk in trait-anxious populations (Chan,
1994; Lovibond & Chan, 1990). Similarly, Tomarken, Mineka, and
Cook (1989) exposed high- and low-fear participants to slides of
snakes, spiders, mushrooms, or flowers, which were followed by
electric shock, a tone, or nothing. Although the association between
the slides and shocks was random, persons with high fear overesti-
mated the relationship between the fear slides and shock, even
though they were relatively accurate with respect to other covariation
estimates. These findings have been replicated with other fear stimuli
(Pury & Mineka, 1997), especially prepared fears (e.g., Davey &
Craigie, 1997).
In addition, judgment bias is observable in interpretations of am-
biguous verbal stimuli, such as tests using homophones (Mathews,
Richards, & Eysenck, 1989), and word selection following ambigu-
ous sentence tasks (Eysenck, Mogg, May, et al., 1991). Anxious
adults and children are more likely to select threatening interpreta-
tions of ambiguous scenes than are healthy controls (Amir, Foa, &
Coles, 1998; Barrett, Rapee, Dadds, & Ryan, 1996; Butler & Math-
ews, 1983; Clark, Salkovskis, Ost, & Breitholtz, 1997). Also, high
trait-anxious individuals demonstrate longer reaction times to am-
biguous sentences that have a possible threat-related interpretation,
compared to low-anxious individuals (MacLeod & Cohen, 1993).

Source of cognitive biases

Attentional biases for threat-relevant information are apparent in all

individuals during periods of high state anxiety, but clearly there is a
disproportionate level of attentional bias in anxious populations. In
addition, it appears as though availability and implicit memory bi-
ases are present as well. Judgment biases—distorted perceptions of
risk—have been implicated in an attempt to understand the develop-

64 Bulletin of the Menninger Clinic

 Cognitive biases in anxiety disorders

ment and maintenance of the attention and memory cognitive biases.

Specifically, Butler and Mathews (1983) theorized that biased esti-
mates of the probability of future threat, and an increased tendency
to interpret ambiguous stimuli as threatening, mediate these accentu-
ated biases in anxious groups (MacLeod & Matthews, 1991). Beck
and Clark (1988) extend this further by proposing that a “danger-
schemata” drives encoding and availability of threat, which in turn
generates perceptions of future risk and elevated distress. As dis-
cussed previously, elevated distress appears to magnify these biases,
thus contributing to a self-perpetuating cycle.
If the source is a danger-schemata, then by implication attentional
bias is a stable trait of anxious individuals, and is responsible for ele-
vated levels of state anxiety. In contrast, Bower (1981) regards atten-
tional biases as more situationally determined. He suggests that acti-
vation of nodes in a person’s emotional network leads to bias in
attention, such that the greater the match between environmental
stimuli and a fear network, the more likely a fear network (and thus
state anxiety) will be activated (Foa & Kozak, 1986; Lang, 1985).
Therefore Bower views anxious mood (or state anxiety) as being pri-
marily responsible for attentional biases, which in turn maintain the
activity of the fear network. It is clear that more research is needed to
determine the nature of the relationship between cognitive biases and
anxious mood. Most likely a reciprocal relationship exists in which
high anxiousness predisposes toward threat biases, and threat biases
in turn elevate anxiety.

Cognitive biases and the process of cognitive-behavioral treatment

Given the presence of these cognitive biases in the anxiety disorders,

it is of utmost importance for us to consider how they affect thera-
peutic process and outcome. Although cognitive-behavioral treat-
ments for anxiety disorders tend to be quite effective, understanding
the way in which cognitive biases influence treatment may enable us
to modify our treatments to minimize the potentially debilitating ef-
fects of these biases in the process of conducting therapy, and thus
maximize the efficacy of treatment. First we consider how cognitive
biases are addressed in cognitive-behavioral therapy for anxiety
disorders.
Targets of cognitive-behavioral therapy
The most common target of cognitive-behavioral interventions is the
set of threat-laden judgments and core beliefs to which biases in at-
tention and memory are linked. The assumption is that by shifting

Vol. 65, No. 1 (Winter 2001) 65

Craske and Pontillo

judgments of risk, biases in attention and memory will “correct,”

and, indeed, there is some support for this notion. For example, at-
tentional biases appear to normalize after successful treatment for
generalized anxiety disorder (Mathews, Mogg, Kentish, & Eysenck,
1995; Mogg, Bradley, Millar, & White, 1995). The same has been
observed in social phobia (Mattia, Heimberg, & Hope, 1993), obses-
sive-compulsive disorder (Foa & McNally, 1986), and specific pho-
bias (Watts, McKenna, Sharrock, & Trezise, 1986). Moreover, in at
least one study (Mogg et al., 1995), degree of change in modified
Stroop latencies for threat-relevant words correlated with change in
perceptions of risk (or anxious thoughts). Nevertheless, there is a
strong need for further testing of treatment effects on attention and
memory biases.
An alternative, less commonly used approach is to target attention
and memory biases directly in the therapeutic endeavor. Attentional
training (Papageorgiou & Wells, 1999) is one such example. This
technique has been shown to be effective for preoccupation with bod-
ily symptoms in hypochondriasis (Papageorgiou & Wells, 1998) and
panic disorder (Wells, 1990). Attentional training aims to reduce per-
severative self-focused processing and increase flexible control over
attention. It is not intended as a strategy for managing acute distress
but instead as a way to change stable qualities of attention. Another
example is our work with biases in recall for prior painful proce-
dures. With children undergoing painful lumbar punctures, we com-
pared an intervention aimed specifically at generating more accurate
recall of coping strategies and responses to prior lumbar punctures
(e.g., correcting negative inflations of prior distress) to an attention
placebo control. The memory-based intervention resulted in less
physiological and psychological distress during subsequent lumbar
punctures (Chen, Zeltzer, Craske, & Katz, 1999). Although these at-
tentional and memory-targeted interventions seem promising, we do
not yet know whether directly targeting processes of memory and at-
tention influences judgments of future risk: Presumably perceived
risk decreases in concert with decreased psychological distress, but
the preceding studies have not directly tested changes in perceived
risk. Nor do we know how well this approach compares to the more
traditional cognitive-behavioral approach, described first, in which
appraisals of risk are targeted directly. Ways in which these ap-
praisals are directly targeted are described next.
Tools for downshifting judgments of risk
Tools for shifting judgments of risk fall into three broad categories:
psychoeducation about errors in judgment, identification of personal

66 Bulletin of the Menninger Clinic

 Cognitive biases in anxiety disorders

errors in judgment, and logical empiricism and data collection to

challenge errors and generate more realistic judgments.
Psychoeducation about cognitive biases and distortions is common
to cognitive-behavioral therapy for anxiety disorders. Understanding
how cognitive errors are inherent to the state of anxiety, and yet can
contribute to the misconstrual of threat and therefore to distress,
contributes to a new metacognition about anxious thinking patterns.
Included in this psychoeducation is information about the common
types of errors in anxious thinking as well as about automaticity of
cognitive biases, which is especially helpful in explaining episodes of
anxiety or fear that seem inexplicable. Taking a personal scientist
perspective enables individuals to become observers of their own cog-
nitive biases and to be willing to consider alternatives to their own
judgments.
Because attentional and expectancy biases can augment estima-
tions of frequency of threat in the environment, accurate self-moni-
toring is essential in providing patients with a more realistic picture
of their world. Thus ongoing written records of anxiety and panic at-
tacks are critical, along with maintaining objective records of auto-
matic thoughts. This type of ongoing record facilitates the identifica-
tion of personal errors in judgment.
Once identified, biases are challenged by having patients act as
empiricists, collecting all available evidence to evaluate their judg-
ments of a given situation. Through practice, patients learn to be
more deliberate and conscious of their own filtering mechanisms, and
they compensate for them by appropriately considering aspects of
their environment that may have been previously ignored because of
lower saliency or threat-relevance. This type of logical empiricism
comes in part from rational consideration of true risk factors, ig-
nored evidence, historical data, alternative explanations, and so on.
Direct exposure to fear-provoking cues provides another avenue for
collecting disconfirming data and generation of more evidence-based
interpretations. For example, persons who believe that they may fall
from a balcony learn that they do not fall when they stand on the bal-
cony, and persons who believe that their obsessions about harming
their children may lead to them to kill their children learn otherwise
as they deliberately induce their obsessive thoughts.

Effect of state anxiety on attempts to downshift risk judgments

As noted, the primary goal of the more usual approach to cognitive-
behavioral therapy is to teach strategies for more accurate assessment
of risk of threat. However, the learning of such strategies may be im-
peded by anxious cognitive biases—the disproportionate allocation

Vol. 65, No. 1 (Winter 2001) 67

Craske and Pontillo

of attentional resources to threat-related stimuli, priming or implicit

memory biases for threat-related material, and biases in judgment of
risk. This is especially likely given that we all selectively attend to
stimuli supporting our beliefs, and ignore or discount information
suggesting otherwise. Moreover, there are reasons to believe that at-
tempts to change perceptions and judgments of risk may be met with
increased resistance as state arousal increases.
The first reason is that, as noted earlier, threat biases in attention
and risk perception appear to be stronger when state anxiety is ele-
vated. For example, Butler and Mathews (1987) demonstrated that
perceived risk of failing an exam was significantly greater the day be-
fore an exam relative to one month before an exam, in direct rela-
tionship with level of state anxiety. Similarly, Chen and Craske
(1998) found that biases in interpretations of ambiguous information
and risk perception before and after an exam, and after grades were
posted for the exam, were predicted by changes in state anxiety.
Moreover, interference effects on the modified Stroop task in spider-
fearful individuals were elevated under conditions of state anxiety
(Chen et al., 1996).*
Conceivably, then, attempts to therapeutically correct judgment
biases may be thwarted by conditions of acute anxiety. For example,
anxiety in the moment may lead the anxious individual to misinter-
pret educational information as threatening (e.g., presentation of a
.000034 probability of a fatal plane crash is interpreted as “there is a
chance that I will die in a plane crash”) or to selectively attend to
negative information within the context of positive therapeutic infor-
mation (e.g., the person overattends to the possibility of negative so-
cial interactions while ignoring the educational message that occa-
sional negative social interactions are normal). Other examples may
include difficulty concentrating on corrective information because of
other worrisome thoughts (e.g., the person is unable to attend to edu-
cation about the normality of intrusive thoughts about causing harm
to others because the person has worrisome thoughts that he or she is
evil and dangerous for having such intrusive thoughts in the first
place), or to have other threat associations primed by exposure to
threat-relevant information, even when that information is presented
in an educational context (e.g., perceived inability to cope with panic

*Failed attempts to demonstrate state elevation of cognitive biases may be

attributable to less sensitive between-subject designs (e.g., MacLeod & Cohen, 1993)
or use of nonspecific, nonpersonally relevant stressors to induce state anxiety (e.g.,
McNally, Riemann, Louro, Lukach, & Kim, 1992).

68 Bulletin of the Menninger Clinic

 Cognitive biases in anxiety disorders

primed by reading educational information about the physiology of

panic attacks).
A second reason is the evidence suggesting that we are more likely
to rely on cognitive “short cuts” under times of stress, and familiar
yet distorted patterns of thinking are short cuts. That is, cognitive re-
sources may be depleted under conditions of stress, and therefore re-
liance on potentially dysfunctional short cuts increases, whereas al-
ternative modes of information processing that require greater
resource expenditure (e.g., attending to data rather than to prior ex-
pectancies) are obstructed (Ford & Kruglanski, 1995; Sweller, 1988;
Wegner & Erber, 1992).
If so, then attempts to downshift judgments of risk may proceed
most effectively under calm conditions, such as those induced by re-
laxation training or by the sense of safety engendered by a knowl-
edgeable and reassuring therapist. Borkovec and Sides (1979) noted
that subjects instructed to relax benefited more from repeated expo-
sure to fear-relevant images than did those not instructed to relax.
They attributed the positive effects of relaxation to enhanced atten-
tion to fear-relevant information and better ability to encode new in-
formation generated by the exposure experience. Similarly, Foa and
Kozak (1986) noted that “persistent high levels of arousal during ex-
posure interfere with encoding and integration of disconfirming in-
formation” (p. 29) about perceived risk and valence.
On the other hand, there is also reason to believe that learning is
maximal under conditions involving a certain level of arousal. That
is, an optimal level of arousal exists for optimal learning, which
varies according to the complexity of the task to be learned (Yerkes
& Dodson, 1908). Moreover, as noted later in this article, mood-
dependency effects imply that new cognitions should be learned and
practiced under a variety of different mood conditions—in relaxed as
well as anxious states. Unfortunately, ways of establishing the opti-
mal level of state anxiety for learning to downshift judgments of risk
are not well formulated. Clearly there is a need for more empirical in-
vestigation of the conditions most conducive to cognitive shifts.
Effect of cognitive biases on exposure to fear-provoking cues
A staple ingredient of cognitive-behavioral therapy for anxiety disor-
ders is systematic, repeated exposure to fear-provoking stimuli. At-
tention, memory, and judgment biases may deter individuals from en-
gaging in exposures to fear-provoking stimuli, may lead to judgment
that such exposures are dangerous, or may contribute to overly nega-
tive self-evaluation of performance during exposure. This may extend
to judgments about fear responses during exposure exercises as a sign

Vol. 65, No. 1 (Winter 2001) 69

Craske and Pontillo

of personal failure (e.g., Wells & Clark, 1997) or as a sign of threat

in the environment. For example, the socially anxious individual will
be inclined to avoid social interactions, place negative interpretations
on ambiguous social cues in social interactions, be primed toward
negative affect and negative interpretations in social situations, over-
attend to perceived signs of negative judgment from others, and judge
personal performance and anxiety negatively in social interactions.
Thus cognitive tools for challenging risk overestimation and overly
negatively valence judgments often are taught in preparation for,
during, and in post-evaluation of exposures to fear-provoking stim-
uli. Interestingly, however, the value of cognitive tools may lie prima-
rily in enabling the undertaking of exposure, given that sustained ex-
posure alone usually generates as much shift in cognition as does the
addition of cognitive restructuring tools to exposure (e.g., Bouchard
et al., 1996; Feske & Chambless, 1995).
As noted, some have suggested that, in contrast to attentional and
implicit memory biases toward threat, there is an avoidance of elabo-
ration of threat-relevant information in anxiety. Also as noted, the
construct of avoidance of elaboration has since been questioned.
Nevertheless the debate continues, and if such avoidance is inherent
to anxiety, then it may impede learning of new associations through
exposure. That is, fear reduction during exposure is believed to de-
pend on the formation of new associations among stimuli, responses,
and meanings (Foa & Kozak, 1986) or among stimuli, meanings, and
contexts (Bouton, 1988). Indirect evidence comes from a detrimental
long-term impact of distraction during in vivo exposure to fear-pro-
voking stimuli (e.g., Craske, Street, & Barlow, 1989). Such distrac-
tion is hypothesized to prevent “functional exposure” to fear stimuli
(Borkovec & Grayson, 1980), and thus ultimately interfere with
emotional processing and eventual fear reduction (Foa & Kozak,
1986). However, overall, results regarding the role of distraction are
equivocal (see Rodriguez & Craske, 1993, for a review). In addition,
strategic efforts to cope by distraction may be unrelated to avoidance
of elaboration and explicit memory.
Mood-dependent memory
Patients often report feeling less anxious during therapy than in the
real-world setting for a variety of reasons. First, fear cues are typi-
cally absent when logically evaluating judgment errors in the office
setting. Second, when fear cues are faced during exposure therapy,
the systematic and predictable nature of exposure assignments fre-
quently results in less distress than during unexpected encounters
with fear stimuli in daily life. Third, therapists often are perceived as

70 Bulletin of the Menninger Clinic

 Cognitive biases in anxiety disorders

“safe,” thus decreasing patients’ levels of anxiety. Consequently,

new, nonfearful information is acquired/learned in a relatively re-
laxed, secure atmosphere. However, mood-dependent effects suggest
that memory for this new information will be more easily retrievable
when in the same mood as when the information was learned. In
daily life, the patient is likely to experience heightened levels of anxi-
ety, a mood state qualitatively different than that in which the infor-
mation was learned. This in turn could result in less recall of the new
information, and thus less success in managing fear. Moreover, given
judgment biases, patients may be likely to remember these highly
salient experiences in which they were unable to successfully imple-
ment therapeutic techniques. To offset these mood-dependent effects,
patients should practice retrieving new information and skills under a
variety of conditions, including those involving heightened levels of
anxiety and arousal.
A related point is the importance of attribution of patients’
progress, and ultimately of their ability to successfully navigate treat-
ment. It is apparent that attribution of self-efficacy during treatment
is often driven by successes and failures experienced throughout
treatment. Increased attention, expectancies, and memory for nega-
tive outcomes likely augment the failures and diminish the successes
that patients experience. In turn, this could lead to distorted beliefs
about their own prognosis or to skepticism about the treatment itself,
both of which could affect outcome.
Overcoming automatic biases
The issue of automatic biases raises two questions. First, can cogni-
tive-behavioral treatments override automatic cognitive biases? A re-
lated question is whether we expect cognitive-behavioral treatment to
erase biases (and, if so, to what extent) or simply teach ways of com-
pensating for biases. To date, only one study has directly targeted
this question by showing that successful treatment of generalized
anxiety disorder resulted in normalization of attentional interference
effects for subliminally presented threat-relevant information (Mogg
et al., 1995). Clearly, replication of this effect is needed, especially in
terms of the long-term results. Nevertheless, given the subliminal na-
ture of the stimulus presentation, the finding reported by Mogg et al.
(1995) suggests that cognitive biases are removed rather than com-
pensated for. However, this does not rule out the possibility that suc-
cessfully treated individuals remain susceptible to a reappearance of
anxious cognitive biases when state anxiety increases under stressful
conditions at some time in the future. Again, further investigation is
sorely needed.

Vol. 65, No. 1 (Winter 2001) 71

Craske and Pontillo

A second question is whether cognitive techniques alone, inde-

pendent of behavioral exposure, can override automatic biases. Mc-
Nally (1994) asserted that cognitive therapies are unlikely to be ef-
fective due to the unconscious and involuntary nature of anxious
cognitive processes. Similarly, Soares and Ohman (1993) concluded
that phobic fear responses, being largely prepared in nature, “are
inaccessible to voluntary control once elicited because they are
rooted in ancient information processing mechanisms outside the
control of conscious intentions and evaluations. Thus, they resist
cognitive conscious manipulations appealing to rational thought”
(pp. 93-94). In contrast, Beck and Clark (1997) argue that “one of
the most effective ways of deactivating the primal threat mode is to
counter it with more elaborative, strategic processing of information
resulting from the activation of the constructive, reflective modes of
thinking” (p. 55). Clearly, cognitive therapy techniques have been
shown to be effective for most of the anxiety disorders. However,
the debate continues as to whether exposure is unavoidably part of
cognitive therapy (e.g., Marks, 1987). Thus the degree to which
change is directly attributable to constructive, reflective modes of
thinking is not clear.

Summary

It appears that cognitive biases in attention and implicit (and possibly

explicit) memory are characteristic of highly anxious populations. In
addition, research has shown that these biases are associated with
distorted judgments of risk, which in turn serve to maintain dispro-
portionate levels of anxiety. It is crucial, then, that one consider these
variables when attempting to understand the development and main-
tenance of anxiety disorders. In this article, we have provided a
framework for understanding the nature of these mechanisms and
how they affect the process and outcome of cognitive-behavioral
treatment. In addition, we have presented several ways of minimizing
the potentially deleterious effects of such biases on the process of cog-
nitive-behavioral therapy. Although vulnerability to biases in judg-
ment and perception may never be completely extinguished through
treatment, we are optimistic that future research will shed further
light on the nature and mechanisms of cognitive change.

References
Amir, N., Foa, E. B., & Coles, M. E. (1998). Automatic activation and strategic
avoidance of threat-relevant information in social phobia. Journal of Abnormal
Psychology, 107(2), 285–290.
Barrett, P. M., Rapee, R. M., Dadds, M. M., & Ryan, S. M. (1996). Family

72 Bulletin of the Menninger Clinic

 Cognitive biases in anxiety disorders

enhancement of cognitive style in anxious and aggressive children. Journal of

Abnormal Child Psychology, 24, 187–203.
Beck, A. T., & Clark, D. A. (1988). Anxiety and depression: An information
processing perspective. Anxiety Research, 1, 23–26.
Beck, A. T., & Clark, D. A. (1997). An information processing model of anxiety:
Automatic and strategic processes. Behaviour Research and Therapy, 35(1),
49–58.
Beck, A. T., Emery, G., & Greenberg, R. L. (1985). Anxiety disorders and phobias: A
cognitive perspective. New York: Basic Books.
Borkovec, T. D., & Grayson, J. B. (1980). Consequence of increasing the functional
impact of internal emotional stimuli. In K. Blanskein, P. Pliner, & J. Policy (Eds.),
Advances in the study of communication and affect: assessment and modification
of emotional behavior (Vol. 6, pp. 117–137). New York: Plenum.
Borkovec, T. D., & Sides, J. K. (1979). The contribution of relaxation and expectancy
to fear reduction via graded, imaginal exposure to feared stimuli. Behaviour
Research and Therapy, 17(6), 529–540.
Bouchard, S., Gauthier, J., Laberge, B., French, D., Pelletier, M. H., & Godbout, C.
(1996). Exposure versus cognitive restructuring in the treatment of panic disorder
with agoraphobia. Behaviour Research and Therapy, 34(3), 213–224.
Bouton, M. (1988). Context and ambiguity in the extinction of emotional learning:
Implications for exposure therapy. Behaviour Research and Therapy, 26(2),
137–149.
Bower, G. H. (1981). Mood and memory. American Psychologist, 36, 129–148.
Bradley, B. P., Mogg, K., & Williams, R. (1994). Implicit and explicit memory for
emotional information in non-clinical populations. Behaviour Research and
Therapy, 32, 65–78.
Broadbent, D. E., & Broadbent, M. (1988). Anxiety and attentional bias: State and
trait. Cognition and Emotion, 2, 165–183.
Butler, G., & Mathews, A. (1983). Cognitive processes in anxiety. Advances in
Behaviour Research and Therapy, 5, 51–62.
Butler, G., & Mathews, A. (1987). Anticipatory anxiety and risk perception.
Cognitive Therapy and Research, 11(5), 551–565.
Chan, C. K. Y. (1994). Expectancy bias in trait anxiety (Doctoral dissertation,
University of New South Wales, Kensington, NSW, Australia, 1994) Dissertation
Abstracts International: Section B: The Sciences & Engineering, 55 (2–B), 624.
Chan, C. K. Y., & Lovibond, P. F. (1996). Expectancy bias in trait anxiety. Journal of
Abnormal Psychology, 105, 637–647.
Chen, E. & Craske, M. G. (1998). Risk perceptions and interpretations of ambiguity
related to anxiety during a stressful event. Cognitive Therapy and Research, 22(2),
137–148.
Chen, E., Lewin, M. R., & Craske, M. G. (1996). Effects of state anxiety on selective
processing of threatening information. Cognition and Emotion, 10(3), 225–240.
Chen, E., Zeltzer, L. K., Craske, M. G., & Katz, E. R. (1999). Alteration of memory
in the reduction of children’s distress during repeated aversive medical procedures.
Journal of Consulting and Clinical Psychology, 67(4), 481–490.
Clark, D. M., Salkovskis, P. M., Ost, L. G., & Breitholtz, E. (1997).
Misinterpretation of body sensations in panic disorder. Journal of Consulting and
Clinical Psychology, 65(2), 203–213.
Craske, M. G., Street, L., & Barlow, D. H. (1989). Instructions to focus upon or
distract from internal cues during exposure treatment of agoraphobic avoidance.
Behaviour Research and Therapy, 27(6), 663–672.

Vol. 65, No. 1 (Winter 2001) 73

Craske and Pontillo

Davey, G. C. L., & Craigie, P. (1997). Manipulation of dangerousness judgments to

fear-relevant stimuli: Effects on a priori UCS expectancy and a posteriori
covariation assessment. Behaviour Research and Therapy, 35(7), 607–617.
Eich, E. (1995). Searching for mood dependent memory. Psychological Science, 6,
67–75.
Eysenck, M. W. (1992). Anxiety: The cognitive perspective. Hillsdale, NJ: Erlbaum.
Eysenck, M. W., & Byrne, A. (1994). Implicit memory bias, explicit memory bias,
and anxiety. Cognition and Emotion, 8, 415–431.
Eysenck, M. W., Mogg, K., May, J., Richards, A., & Mathews, A. (1991). Bias in
interpretation of ambiguous sentences related to threat in anxiety. Journal of
Abnormal Psychology, 100, 144–150.
Feske, U., & Chambless, D. L. (1995). Cognitive behavioral versus exposure only
treatment for social phobia: A meta-analysis. Behavior Therapy, 26(4), 695–720.
Foa, E. B., Franklin, M. E., Perry, K. J., & Herbert, J. D. (1996). Cognitive biases in
generalized social phobia. Journal of Abnormal Psychology, 105, 433–439.
Foa, E. B., & Kozak, M. J. (1986). Emotional processing of fear: Exposure to
corrective information. Psychological Bulletin, 99, 20–35.
Foa, E. B., & McNally, R. (1986). Sensitivity to feared stimuli in obsessive-
compulsives: A dichotic listening analysis. Cognitive Therapy and Research, 10,
477–485.
Foa, E. B, McNally, R., & Murdock, T. B. (1989). Anxious mood and memory.
Behaviour Research and Therapy. 27(2), 141–147.
Ford, T. E., & Kruglanski, A. W. (1995). Effects of epistemic motivations on the use
of accessible constructs in social judgment. Personality and Social Psychology
Bulletin, 21(9), 950–962.
Graf, P., & Mandler, G. (1984). Activation makes words more accessible, but not
necessarily more retrievable. Journal of Verbal Learning and Verbal Behavior,
23(5), 553–568.
Graf, P., & Masson, M. E. J. (Eds.). (1993). Implicit memory: New directions in
cognition, development, and neuropsychology. Hillsdale, NJ: Erlbaum.
Lang, A., Craske, M. G., Brown, M., & Ghaneian, A. (in press). Fear-related state
dependent memory. Cognitive Therapy and Research.
Lang, P. J. (1985). The cognitive psychophysiology of emotion: Fear and anxiety. In
A. H. Tuma & J. D. Maser (Eds.), Anxiety and the anxiety disorders (pp.
131–170). Hillsdale, NJ: Erlbaum.
Logan, A. C., & Goetsch, V. L. (1993). Attention to external threat cues in anxiety
states. Clinical Psychology Review, 13(6), 541–559.
Lovibond, P.F., & Chan, C. K. Y. (1990). Threat appraisal and trait anxiety. In N.
McNaughton & G. Andrews (Eds.), Anxiety (pp. 160–168). Dunedin, New
Zealand: University of Otago Press.
Lundh, L. G., Czyzykow, S., & Ost, L. G. (1997). Explicit and implicit memory bias
in panic disorder with agoraphobia. Behaviour Research and Therapy, 35(11),
1003–1014.
MacLeod, C., & Cohen, I. L. (1993). Anxiety and the interpretation of ambiguity: A
text comprehension study. Journal of Abnormal Psychology, 102(2), 238–247.
MacLeod, C., & Hagan, R. (1992). Individual differences in the selective processing
of threatening information and emotional responses to a stressful life event.
Behaviour Research and Therapy, 30(2), 151–161.
MacLeod, C., & Mathews, A. (1988). Anxiety and the allocation of attention to
threat. Quarterly Journal of Experimental Psychology: Human Experimental
Psychology, 38, 659–670.

74 Bulletin of the Menninger Clinic

 Cognitive biases in anxiety disorders

MacLeod, C., & Mathews, A. (1991). Biased cognitive operations in anxiety:

Accessibility of information, or assignment of processing priorities? Behaviour
Research and Therapy, 29, 599–610.
MacLeod, C., & McLaughlin, K. (1995). Implicit and explicit memory bias in
anxiety: A conceptual replication. Behaviour Research and Therapy, 33(1),
1–14.
Marks, I. M. (1987). Behavioral aspects of panic disorder. American Journal of
Psychiatry, 144(9), 1160–1165.
Mathews, A., & MacLeod, C. (1985). Selective processing of threat cues in anxiety
states. Behaviour Research and Therapy, 23, 563–569.
Mathews, A., & MacLeod, C. (1986). Discrimination of threat cues without
awareness in anxiety states. Journal of Abnormal Psychology, 95, 131–138.
Mathews, A., Mogg, K., Kentish, J., & Eysenck, M. (1995). Effect of psychological
treatment on cognitive bias in generalized anxiety disorder. Behaviour Research
and Therapy, 33(3), 293–303.
Mathews, A., Mogg, K., May, J., & Eysenck, M. (1989). Implicit and explicit
memory bias in anxiety. Journal of Abnormal Psychology, 98(3), 236–240.
Mathews, A., Richards, A., & Eysenck, M. W. (1989). The interpretation of
homophones related to threat in anxiety states. Journal of Abnormal Psychology,
98, 31–34.
Mattia, J., Heimberg, R. G., & Hope, D. A. (1993). The revised Stroop color-naming
task in social phobics. Behaviour Research and Therapy, 31(3), 305–313.
McNally, R. J. (1994). Cognitive bias in panic disorder. Current Directions in
Psychological Science, 3(4), 129–132.
McNally, R. J., Foa, E. B., & Donnell, C. D. (1989). Memory bias for anxiety
information in patients with panic disorder. Cognition and Emotion, 3(1), 27–44.
McNally, R. J., Metzger, L. J., Lasko, N. B., Clancy, S. A., & Pitman, R. K. (1998).
Directed forgetting of trauma cues in adult survivors of childhood sexual abuse
with and without posttraumatic stress disorder. Journal of Abnormal Psychology,
107(4), 596–601.
McNally, R. J., Otto, M. W., Yap, L., Pollack, M. H., & Hornig, C. D. (1999). Is
panic disorder linked to cognitive avoidance of threatening information? Journal
of Anxiety Disorders, 13(4), 335–348.
McNally, R. J., Riemann, B. C., Louro, C. E., Lukach, B. M., & Kim, E. (1992).
Cognitive processing of emotional information in panic disorder. Behaviour
Research and Therapy, 30(2), 143–149.
Menzies, R. G., & Clarke, J. C. (1995). Danger expectancies and insight in
acrophobia. Behaviour Research and Therapy, 33, 215–221.
Mineka, S., & Sutton, S. K. (1992). Cognitive biases and the emotional disorders.
Psychological Science, 3(1), 65–69.
Mogg, K., Bradley, B. P., Millar, N., & White, J. (1995). A follow-up study of
cognitive bias in generalized anxiety disorder. Behaviour Research and Therapy,
33(8), 927–935.
Mogg, K., Bradley, B. P., & Williams, R. (1995). Attentional bias in anxiety and
depression: The role of awareness. British Journal of Clinical Psychology, 34(1),
17–36.
Mogg, K., Gardiner, J. M., Stavrou, A., & Golombok, S. (1992). Recollective
experience and recognition memory for threat in clinical anxiety states. Bulletin
of the Psychonomic Society, 30(2), 109–112.
Mogg, K., & Mathews, A. (1990). Is there a self-referent mood-congruent recall bias
in anxiety? Behaviour Research and Therapy, 28, 91–92.

Vol. 65, No. 1 (Winter 2001) 75

Craske and Pontillo

Mogg, K., Mathews, A., & Eysenck, M. (1992). Attentional bias to threat in clinical
anxiety states. Cognition and Emotion, 6, 149–159.
Mogg, K., Mathews, A., & Weinman, J. (1987). Memory bias in clinical anxiety.
Journal of Abnormal Psychology, 96(2), 94–98.
Mogg, K., Mathews, A., & Weinman, J. (1989). Selective processing of threat cues
in anxiety states: A replication. Behaviour Research and Therapy, 27(4),
317–323.
Nugent, K., & Mineka, S. (1994). The effect of high and low trait anxiety on implicit
and explicit memory tasks. Cognition and Emotion, 8(2), 147–163.
Papageorgiou, C., & Wells, A. (1998). Effects of attention training on
hypochondriasis: A brief case series. Psychological Medicine, 28(1), 193–200.
Papageourgiou, C., & Wells, A. (1999). The observer perspective: Biased imagery in
social phobia, agoraphobia, and blood/injury phobia. Behaviour Research and
Therapy, 37(7), 653–658.
Pury, C. L. S., & Mineka, S. (1997). Covariation bias for blood-injury stimuli and
aversive outcomes. Behaviour Research and Therapy, 35, 35–47.
Rapee, R. M., McCallum, S. L., Melville, L. F., Ravenscroft, H., & Rodney, J. M.
(1994). Memory bias in social phobia. Behaviour Research and Therapy, 32,
89–99.
Richards, A., & French, C. C. (1991). Effects of encoding and anxiety on implicit
and explicit memory performance. Personality and Individual Differences, 12(2),
131–139.
Richards, A., & French, C. C. (1992). An anxiety-related bias in semantic activation
when processing threat/neutral homographs. Quarterly Journal of Experimental
Psychology: Human Experimental Psychology, 45(3), 503–525.
Richards, A., & Millwood, B. (1989). Colour-identification of differentially valenced
words in anxiety. Cognition and Emotion, 3(2), 171–176.
Rieman, B. C., & McNally, R. J. (1995). Cognitive processing of personally relevant
information. Cognition and Emotion, 9(4), 325–340.
Rodriguez, B. I., & Craske, M. G. (1993). The effects of distraction during exposure
to phobic stimuli. Behaviour Research and Therapy, 31(6), 549–558.
Soares, J. F., & Ohman, A. (1993). Preattentive processing, preparedness, and
phobias. Behaviour Research and Therapy, 31, 87–96.
Sweller, J. (1988). Cognitive load during problem solving: Effects on learning.
Cognitive Science, 12(2), 257–285.
Tata, P. R., Leibowitz, J. A., Prunty, M. J., Cameron, M., & Pickering, A. D. (1996).
Attentional bias in obsessive-compulsive disorder. Behaviour Research and
Therapy, 34, 53–60.
Taylor, S., & Rachman, S. (1994). Stimulus estimation and the overprediction of fear.
British Journal of Clinical Psychology, 33, 173–181.
Tomarken, A. J., Mineka, S., & Cook, M. (1989). Fear-relevant selective associations
and covariation bias. Journal of Abnormal Psychology, 98(4), 381–394.
Watts, F. N., McKenna, F., Sharrock, R., & Trezise, L. (1986). Colour naming of
phobia-related words. British Journal of Psychology, 77, 97–108.
Wegner, D. M., & Erber, R. (1992). The hyperaccessibility of suppressed thoughts.
Journal of Personality and Social Psychology, 63(6), 903–912.
Wells, A. (1990). Panic disorder in association with relaxation induced anxiety: An
attentional training approach to treatment. Behavior Therapy, 21(3), 273–280.
Wells, A., & Clark, D. M. (1997). Cognitive therapy for anxiety disorders. In L. J.
Dickstein & M. B. Riba (Eds.), American Psychiatric Press review of psychiatry
(Vol. 16, pp. 9–43). Washington, DC: American Psychiatric Press.

76 Bulletin of the Menninger Clinic

 Cognitive biases in anxiety disorders

Williams, J. M. G., Watts, F. N., MacLeod, C., & Mathews, A. (1988). Cognitive
psychology and emotional disorders. New York: Wiley.
Yerkes, R. M., & Dodson, J. D. (1908). The relation of strength of stimulus to
rapidity of habit-formation. Journal of Comparative and Neurological
Psychology, 18, 459–482.

Vol. 65, No. 1 (Winter 2001) 77