Personality and Individual Differences 50 (2011) 955960

Contents lists available at ScienceDirect

Personality and Individual Differences

journal homepage: www.elsevier.com/locate/paid

New perspectives in attentional control theory

Michael W. Eysenck a,, Nazanin Derakshan b
a
b

Department of Psychology, Roehampton University, Whitelands College, London SW15 4JD, UK

Birkbeck University of London, UK

a r t i c l e

i n f o

Article history:
Received 20 April 2010
Received in revised form 11 August 2010
Accepted 18 August 2010
Available online 12 October 2010
Keywords:
Anxiety
Attentional control
Distractibility
Processing efciency
Performance effectiveness

a b s t r a c t
There have been several theoretical attempts to explain the effects of anxiety on cognitive performance.
According to attentional control theory, anxiety impairs the efciency of two executive functions (the
inhibition and shifting functions). Another major theoretical assumption is that anxiety impairs performance effectiveness (the quality of performance) to a lesser extent than processing efciency (the relationship between performance effectiveness and effort or use of processing resources). However, there
may be conditions (e.g., prior presentation of threat-related stimuli) in which that assumption is not
applicable. The extensive recent research (including several cognitive neuroscience studies) of direct relevance to the theory is discussed, and suggestions are made for maximizing the value of future cognitive
neuroscience research. Finally, attentional control theory is developed to explicate the relationship
between anxiety and motivation. Implications for theoretical predictions and alternative theoretical
accounts are discussed.
2010 Elsevier Ltd. All rights reserved.

1. Introduction
This article is concerned with the effects of individual differences in anxiety on cognitive performance. The emphasis is mainly
on anxiety as a personality dimension (i.e., trait anxiety or test anxiety), but the effects of transient anxiety (i.e., state anxiety) are also
considered. There is plentiful evidence that anxiety (whether
regarded as a personality dimension or as an emotional state) is
associated with performance impairments on numerous tasks. A
meta-analysis based on hundreds of studies revealed an overall
correlation of 0.29 between test anxiety and academic aptitude
or achievement (Hembree, 1988).
There have been numerous attempts to provide a theoretical
explanation for the adverse effects of anxiety on performance.
However, we will focus on one particular theoretical approach that
has evolved over time. The original statement of the theory was by
Eysenck (1979). This was followed by processing efciency theory
(Eysenck & Calvo, 1992), and more recently by attentional control
theory (Derakshan & Eysenck, 2009; Eysenck, Derakshan, Santos, &
Calvo, 2007). It would be superuous to describe in detail the
development of the theory. Instead, we will consider only the major theoretical hypotheses incorporated within processing efciency theory and attentional control theory (the additional
hypotheses associated with attentional control theory are discussed at length in Eysenck et al.).

Corresponding author.
E-mail address: m.eysenck@rhul.ac.uk (M.W. Eysenck).
0191-8869/$ - see front matter 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.paid.2010.08.019

The rst major hypothesis forms an important part of processing efciency theory (Eysenck & Calvo, 1992): anxiety impairs the
efciency of the central executive, which is an attention-like, limited capacity component of the working memory model put forward by Baddeley (1986). In contrast, it was assumed that there
are only modest effects of anxiety on the other two components
of the original model: (1) the phonological loop (used to rehearse
verbal material and to store it briey) and (2) the visuo-spatial
sketchpad (used to process and store transiently visual and spatial
information).
Much evidence provides support for the rst hypothesis in
terms of the effects of anxiety on the central executive (see
Derakshan & Eysenck, 2009, for a review). However, there are very
few studies in which the effects of anxiety on all three components
of the working memory model have been compared directly in a
single experiment. One such study was carried out by Eysenck,
Payne, and Derakshan (2005). Individuals high and low in trait
anxiety performed the Corsi Blocks Test concurrently with a secondary task involving the central executive, the phonological loop,
or the visuo-spatial sketchpad. Performance on the Corsi Blocks
Test was impaired by high trait anxiety when the secondary task
involved use of the central executive but not when it involved
use of the phonological loop or the visuo-spatial sketchpad. These
ndings suggested that high anxiety only impaired the functioning
of the central executive.
Christopher and MacDonald (2005) used a different approach to
the same issue. Their participants performed a series of tasks designed to assess different components of the working memory system. The ndings closely resembled those of Eysenck et al. (2005):

956

M.W. Eysenck, N. Derakshan / Personality and Individual Differences 50 (2011) 955960

high trait anxiety only impaired performance on those tasks

involving the central executive. Walkenhorst and Crowe (2010)
used the same general approach as Christopher and MacDonald.
They also failed to nd any signicant effects of trait anxiety on
tasks involving the phonological loop or the visuo-spatial sketchpad. However, in contrast to the ndings of Christopher and MacDonald, they also reported non-signicant effects of anxiety on
tasks involving the central executive.
The second and third major hypotheses are novel to attentional
control theory but developed out of the assumption that anxiety
impairs the functioning of the central executive. In short, there is
accumulating evidence that various executive functions are associated with the central executive. For example, Miyake et al. (2000)
asked their participants to perform numerous executive tasks, and
then performed latent-variable analysis to identify the main
underlying executive functions. They identied partially independent inhibition, shifting, and updating functions. The inhibition
function (subsequently claried by Friedman and Miyake (2004))
prevents task-irrelevant stimuli and responses from disrupting
performance. The shifting function is used to allocate attention in
a exible and optimal way to the task stimulus or stimuli that
are currently most relevant. The updating function is used to update and monitor the information currently within working memory. This function is important for various short-term memory
tasks.
The second hypothesis is that anxiety impairs the functioning of
the inhibition function, and the third hypothesis is that anxiety impairs the functioning of the shifting function. The overarching
assumption is that anxiety impairs attentional control whether
that control is negative (inhibition function) or positive (shifting
function).
The fourth major hypothesis is based on the distinction between processing efciency and performance effectiveness, and is
incorporated in all versions of the theory from Eysenck (1979) onwards. Performance effectiveness can be dened as the quality of
performance (e.g., the percentage of correct task responses). Processing efciency is dened by the relationship between performance effectiveness and the use of resources or effort. More
specically, processing efciency is high when performance effectiveness is high and use of resources is low and it is low when performance effectiveness is low but use of resources is high. The
crucial hypothesis is that anxiety will typically impair processing
efciency to a greater extent than performance effectiveness.
Bishop (2009) agreed with the assumption that anxiety is associated with a broad impairment of attentional control. However,
she offered an alternative explanation of the underlying processes.
According to her account, individuals high in trait anxiety often exhibit an impoverished recruitment of attentional control mechanisms centered on the dorsolateral prefrontal cortex. This differs
from the prediction stemming from attentional control theory,
namely, that high anxiety will often be associated with increased
activation of brain areas (e.g., dorsolateral prefrontal cortex) associated with attentional control.
We have already discussed Hypothesis 1. As a result, we will
focus in what follows on the remaining hypotheses.
Hypothesis 2. Anxiety impairs the inhibition function.
The inhibition function is used to resist interference from taskirrelevant stimuli and responses. There is considerable evidence
that anxiety impairs the functioning of the inhibition function
(see Derakshan & Eysenck, 2009, for a review). Most of the research
included in that review involved paradigms in which the conditions varied in terms of the presence and/or nature of distracting
stimuli. The typical nding was that high-anxious individuals were
more susceptible to distraction than were low-anxious individuals,

a nding that has been replicated in more recent research (e.g.,

Pacheco-Ungietti, Acosta, Callejas, & Lupianez, 2010; PachecoUngietti, Lupianez, & Acosta, 2009).
It is important for research in this area to use conceptually simple tasks that are as process pure as possible so that the ndings
obtained are interpretable. The antisaccade task was identied by
Miyake et al. (2000) as such a task. On this task, a visual cue is presented to the left or the right of the xation point, and the instructions are to make an eye movement to the opposite side of the
visual cue as rapidly as possible. The latency of the rst saccade
to the correct side is one of the main dependent variables of interest. There is also a control task (the pro-saccade task), in which the
instructions are to xate the cue when it appears.
According to attentional control theory, adverse effects of anxiety in terms of latency of the rst correct saccade should be present with the antisaccade task but not the pro-saccade task.
Derakshan, Ansari, Hansard, Shoker, and Eysenck (2009) obtained
the predicted pattern of ndings in their rst experiment in which
the cue consisted of a neutral oval shape. In their second experiment, Derakshan et al. used angry, happy, and neutral faces as
cues. They obtained similar ndings to their rst experiment, with
the adverse effects of anxiety on the antisaccade task being greatest with angry cues.
Two other recent studies have investigated the effects of
anxiety on the antisaccade task using emotional cues. Garner,
Ainsworth, Gould, Gardner, and Baldwin (2009) used negative
and neutral picture cues. The high-anxious group made signicantly more eye-movement errors than the low-anxious group
on antisaccade trials regardless of cue type. Wieser, Pauli, and
Muhlberger (2009) used happy, fearful, sad, and neutral facial
expressions as cues in their study. Individuals high in social anxiety made more eye-movement errors than controls on antisaccade
trials for all facial expressions.
Ansari and Derakshan (in press) successfully replicated the previous ndings of Derakshan et al. (2009) using neutral cues. They
also used additional conditions to explain more fully why the latency of the rst correct eye saccade is greater for high-anxious
than for low-anxious individuals. They discovered that this effect
is due predominantly to the negative effects of anxiety on inhibitory control rather than on volitional action generation. This
strengthens the argument that performance on the antisaccade
task is relevant to attentional control theory.
We have discussed convincing evidence that anxiety is often
associated with increased susceptibility to distraction and thus impaired efciency of the inhibition function. As mentioned earlier,
there are two contrasting theoretical accounts of this inefciency
when the inhibition function is required. According to attentional
control theory, the inefciency of high-anxious individuals often
manifests itself in greater activation in brain areas associated with
attentional control in distraction conditions. In contrast, Bishop
(2009) argued that the inefciency of high-anxious individuals in
distraction conditions is often due to a partial failure to engage
attentional control mechanisms (revealed by reduced activation
in brain regions associated with attentional control).
The evidence obtained by Bishop (2009) using functional magnetic resonance imaging (fMRI) suggests that each approach is correct under certain conditions. A horizontal string of six letters was
presented on each trial with the target letter being presented six
times in the string (low perceptual load) or once (high perceptual
load). The task was to decide which target letter (N or X) was presented. In addition, a task-irrelevant letter congruent or incongruent with the target letter was presented slightly above or below the
letter string. In the high perceptual load-condition, there was no
difference in performance (target detection time) between the congruent and incongruent distractor conditions for groups low or
high in trait anxiety. However, only individuals high in trait

M.W. Eysenck, N. Derakshan / Personality and Individual Differences 50 (2011) 955960

anxiety had greater activation in the left dorsolateral prefrontal

cortex (associated with attentional control) in the incongruent
than the congruent condition. These ndings are consistent with
attentional control theory. Even though there were no effects of
trait anxiety on performance with incongruent distractors, individuals high in trait anxiety had additional activation in brain areas
associated with attentional control.
The ndings were very different with low perceptual load. In
this condition, only the high-anxious group had signicantly worse
performance with incongruent distractors than with congruent
ones. In addition, high-anxious participants tended to have reduced activation in the left dorsolateral prefrontal cortex with
incongruent distractors than with congruent ones, whereas lowanxious ones tended to have increased activation. Thus, the impaired performance of high-anxious participants in the presence
of incongruent distractors is likely to be due to a failure to use
attentional control mechanisms when this was necessary for good
task performance.
An important limitation with fMRI is that it has relatively poor
temporal resolution. The inability to pinpoint precisely when increased brain activation occurred complicates the task of working
out which cognitive processes are associated with changing levels
of brain activation. In contrast, event-related potentials (ERPs) provide excellent temporal resolution. Ansari and Derakshan measured ERPs to the onset of the to-be-inhibited cue in an
antisaccade task (see above) in low- and high-anxious groups.
Behavioral results replicated previous ndings of longer antisaccade latencies in high-anxious compared with low-anxious individuals. Analysis of ERPs showed that longer antisaccade
latencies in the high-anxious individuals were associated with
lower fronto-parietal negativity in the period immediately prior
to the onset of the to-be-inhibited cue. Greater negativity prior
to the initiation of correct antisaccades probably reects a cortical
inhibition mechanism involving fronto-parietal control regions
(e.g., Everling, Matthews, & Flohr, 2001; Ford, Goltz, Brown, & Everling, 2005).
There are two more studies relevant to the issue of the effects of
anxiety on the efciency of the inhibition function. Righi, Mecacci,
and Viggiano (2009) used event-related potentials (ERPs) in a
study on the Sustained Attention to Response Task (SART). On this
task, participants are presented with a series of digits. They are instructed to respond to nearly every digit (Go trials) but to inhibit a
response to the digit 3 (NoGo trials). There were no effects of trait
anxiety on performance on the SART. However, individuals high in
trait anxiety had a larger N2 response than those low in trait anxiety on NoGo trials but not on Go trials. This N2 response was
found in anterior sites and may well reect inhibitory processes
within frontal networks. Thus, high anxiety did not affect performance effectiveness but did impair processing efciency when
the inhibition function was required.
Savostyanov et al. (2009) used a measure of EEG desynchronization with excellent temporal resolution. Participants decided as
rapidly as possible the category to which each of a series of words
belonged (the Go condition). There was also a Stop condition
involving the inhibition function in which a signal indicated that
the participants should inhibit their response on that trial.
Savostyanov et al. (2009) did not nd any effects of trait anxiety
on performance effectiveness in the Go or Stop conditions. However, the high-anxious participants had considerably more EEG
desynchronization (indicative of processing effort and use of resources) than the low-anxious ones in both conditions. Thus, high
anxiety was associated with impaired processing efciency of the
type predicted by attentional control theory. Of greatest theoretical
signicance were the detailed ndings in the Stop condition based
on EEG desynchronization in the time period before and after the
warning signal indicating that the response should be inhibited.

957

The greater EEG desynchronization associated with high trait anxiety was almost exclusively limited to the time after the warning
signal, a time during which participants were making use of the
inhibition function. Thus, these ndings support the hypothesis
that anxiety impairs the efciency of the inhibition function.
Hypothesis 3. Anxiety impairs the shifting function.
The effects of anxiety on the shifting function can be assessed
by tracking eye movements on tasks on which it is possible to
specify how visual attention should shift over time. Wilson, Vine,
and Wood (2009) reported a study representing an approximation
to this state of affairs. They found that high anxiety was associated
with impaired use of the shifting function and attentional control
on a basketball shooting task as revealed by the pattern of eye
movements.
The task-switching paradigm provides a relatively direct assessment of the shifting function (Miyake et al., 2000). The basic paradigm involves two conditions in each of which participants
perform the same two tasks (A and B). In the control condition,
each block of trials is devoted to only one task. In the experimental
condition, each block consists of a mixture of trials on task A and
task B. The slowing and/or greater number of errors in the experimental than in the control condition provides an assessment of the
shifting function.
Other tasks provide an approximate measure of the shifting
function. These tasks include the Wisconsin Card Sorting Test
(WCST) and the Comprehensive Trail Making Test (CTMT). Some
evidence based on these tasks supports the hypothesis. Orem, Petrac, and Bedwell (2008) found that highly stressed (and so presumably anxious) individuals performed signicantly more slowly than
less stressed and anxious individuals on Trial 5 of the CTMT, which
involves set switching. Goodwin and Sher (1992) found that highanxious individuals made more errors and took longer to complete
the WCST than did low-anxious individuals. Caselli, Reiman, Hentz,
Osbourne, and Alexander (2004) found that anxious personality
was associated with more errors on the WCST.
Much recent research has involved use of various versions of the
task-switching paradigm. Derakshan, Smyth, and Eysenck (2009)
used pairs of tasks (multiplication and division or addition and
subtraction). There was a highly signicant interaction between
anxiety and task-switching. In this interaction, the high-anxious
participants were considerably slower in the task-switching condition than in the control condition, but there was no difference in
performance speed in the two conditions for low-anxious
participants.
Ansari, Derakshan, and Richards (2008) made use of the mixed
antisaccade and pro-saccade task. They compared performance in
the standard condition (entire blocks involving only one task)
and in a task-switching condition (each block involved a mixture
of pro-saccade and antisaccade trials). Ansari et al. replicated previous ndings in which the latency of the rst correct saccade on
the antisaccade task was faster in the task-switching than in the
control condition. However, they reported the additional nding
that this paradoxical improvement in the task-switching condition
was restricted to low-anxious participants. This nding suggests
that high-anxious participants use the shifting function less efciently than do low-anxious ones.
In a recent investigation using ERPs, Ansari and Derakshan
examined the shifting function in low- and high-anxious groups
in a mixed antisaccade task in which the color of a xation cross
at trial onset instructed participants to make either an anti- or a
pro-saccade. The time between the offset of the xation cross
and the to-be-inhibited cue was manipulated. This was the anticipatory period prior to the execution of the correct saccade. There
were short, medium, and long anticipatory periods. Behavioral

958

M.W. Eysenck, N. Derakshan / Personality and Individual Differences 50 (2011) 955960

ndings showed that high-anxious individuals incurred greater

switch costs than low-anxious ones on trials with a short anticipatory period, but this group difference disappeared on trials with a
longer anticipatory period. Analysis of ERPs to the onset of the
instructional cue showed that the switch cost in high-anxious individuals was associated with lower frontal and fronto-central activity compared with low-anxious ones. Most interestingly, however,
analysis of the ERPs in the longer anticipatory period revealed that
high-anxious individuals had greater slow wave negativity (CNV:
contingent negative variation) at frontal and fronto-central sites
indicative of a greater allocation of compensatory resources to processes in anticipation of the upcoming cue (cf., Rosler, Heil, & Roder, 1997). CNV is examined as a neural marker of resource
allocation and cognitive effort (e.g., Jennings & van der Molen,
2005). These ndings indicate that anxiety impairs prefrontal control functions associated with the shifting function, but this
impairment is associated with greater compensatory effort to
maintain performance quality.
Johnson (2009) used a task-shifting paradigm in which one task
involved emotional processing based on facial expressions and the
other task was neutral. Individuals high in trait anxiety took significantly longer than those low in trait anxiety to switch from the
neutral to the emotional task, thus providing partial support for
the hypothesis.
Kofman, Meiran, Greenberg, Balas, and Cohen (2006) used a
spatial task-switching task. There were two groups of participants:
one group was more stressed and anxious than the other because
they were tested shortly before the start of an examination period.
The two groups were non-signicantly different in terms of
switching costs.
According to attentional control theory, high-anxious individuals typically exhibit inefcient use of the shifting function even
when there are non-signicant effects of anxiety on performance.
Santos, Wall, and Eysenck (submitted for publication) considered
the effects of anxiety on task-switching using three simple tasks
in blocks of trials that involved high switching, low switching, or
no switching. There were no effects of anxiety on speed or accuracy
or performance. However, fMRI ndings indicated that the increase
in brain activation associated with switching was greater for highanxious than for low-anxious individuals, and that the brain areas
involved (principally BA9/46) corresponded to those associated
with task-switching (Wager, Jonides, & Reading, 2004). Thus, anxiety impaired the efciency of the shifting function even though
performance was not signicantly affected.
In sum, the evidence predominantly supports the hypothesis.
There are negative effects of high anxiety with regard to use of
the shifting function. In studies in which there is only behavioral
evidence, these negative effects are often observed with respect
to performance effectiveness. When measures of effectiveness
and efciency are both available, high anxiety typically has a greater adverse effect on efciency than on effectiveness.
There is an important caveat with respect to most of the ndings. The task-switching paradigm probably provides the most direct way of assessing the effects of anxiety on the shifting function.
However, even ndings from this paradigm are typically difcult to
interpret unequivocally. The reason for this is that several factors
jointly determine the costs associated with task-switching (see
Monsell, 2003, for a review).
Hypothesis 4. Anxiety impairs processing efciency more than
performance effectiveness.
Numerous behavioral studies provide indirect support for the
hypothesis that anxiety impairs processing efciency more than
performance effectiveness (see Eysenck et al., 2007, for a review).
An important alternative approach used increasingly in recent

years has involved considering processing efciency by using various techniques for assessing brain activity (e.g., fMRI; ERPs). Several studies based on that approach (Bishop, 2009; Righi et al.,
2009; Santos et al., submitted for publication; Savostyanov et al.,
2009) have shown that high anxiety can be associated with greater
brain activity than low anxiety even when there are no effects of
anxiety on performance. Those studies are of particular theoretical
relevance because they involved tasks varying in their demands on
the inhibition or shifting function. There are other studies (to
which we now turn) providing additional evidence that anxiety
is often associated with impaired processing efciency.
Fales et al. (2008) used the 3-back task on which participants
indicated whether a given word was the same as the one displayed
three words back. There was a non-signicant effect of anxiety on
performance. However, high-anxious participants had greater
transient activation in brain areas (e.g., dorsolateral and ventrolateral prefrontal cortex) associated with attentional control. These
ndings suggest that anxiety impaired processing efciency but
not performance effectiveness.
Telzer et al. (2008) studied attentional bias as a function of trait
anxiety. In the conditions of interest, two faces (one angry and one
neutral) were presented. The faces were followed by a visual probe
and participants indicated as rapidly as possible whether the probe
was on the left or the right. Of particular interest was brain activation on angry-incongruent trials (angry face and probe on different
sides) and on angry-congruent trials (angry face and probe on the
same side). The increase in dorsolateral prefrontal cortex activation
on incongruent trials compared to congruent ones was greater for
high-anxious than for low-anxious individuals. These ndings suggest that the high-anxious participants required greater use of the
inhibition function than low-anxious ones to disengage from the
processing of angry faces.
2. New theoretical and experimental directions
The most fundamental overarching assumption within processing efciency theory and attentional control theory is that behavioral evidence concerning performance effectiveness often
provides very indirect evidence concerning internal processes.
There is plentiful evidence to support that assumption. For example, in several studies there were no effects of anxiety on performance, but signicant effects of anxiety on neuroimaging or EEG
measures suggested that anxiety was affecting various cognitive
processes.
2.1. Cognitive neuroscience
In principle, cognitive neuroscience research provides an extremely useful way of comparing processing efciency in high-anxious and low-anxious individuals. However, various requirements
need to be satised for such research to be maximally informative
with respect to attentional control theory. First, it is important to
compare brain activity in two or more conditions differing primarily in terms of the involvement of the inhibition function or the
shifting function. More specically, the baseline condition should
be one in which there is little or no use of a given executive function and the other condition one necessitating use of that function.
Second, the brain areas that differ between high-anxious and
low-anxious groups should be predicted a priori on the basis of theoretical considerations. Sufcient is known of the main brain areas
associated with specic executive functions for that to be feasible,
although it is important not to exaggerate the amount of functional
specialization within the brain.
Third, the techniques used should have good spatial and temporal resolution. The reason for this is to ensure that the differences

M.W. Eysenck, N. Derakshan / Personality and Individual Differences 50 (2011) 955960

in processing between high-anxious and low-anxious individuals

can be identied as precisely as possible. For example, suppose
that high-anxious individuals performing a given task show a
greater increase in dorsolateral prefrontal cortex activation assessed by fMRI compared to a baseline condition than low-anxious
ones. This might reect greater use of attentional control mechanisms by high-anxious individuals. However, the relatively poor
temporal resolution of fMRI means it is not possible to identify precisely the stage of task processing associated with the difference in
dorsolateral prefrontal cortex activation.
Fourth, it is often implicitly assumed that high-anxious and
low-anxious individuals use the same cognitive processes when
performing any given task. On that assumption, differences between high-anxious and low-anxious individuals in brain activation reect purely quantitative differences in the involvement of
a given cognitive process. However, it is probable that there are
qualitative differences in the cognitive strategies used by highanxious and low-anxious individuals on many tasks (e.g., Tohill &
Holyoak, 2000). As a consequence, it is important to use tasks on
which nearly all participants use the same cognitive strategy. If
that is not the case, it becomes extremely difcult to interpret
group differences in patterns of brain activation.
2.2. Processing efciency: role of motivation
As discussed earlier, high-anxious individuals can exhibit greater processing inefciency than low-anxious ones for two reasons.
According to attentional control theory, high-anxious individuals
often use compensatory strategies such as enhanced effort and
use of processing resources to achieve a reasonable level of performance effectiveness. Processing inefciency within this theoretical
approach is indexed by a smaller ratio of performance effectiveness
to use of processing resources for high-anxious than for low-anxious individuals. According to Bishop (2009), in contrast, processing inefciency in high-anxious individuals often takes the form
of decient use of attentional control mechanisms based on the
dorsolateral prefrontal cortex.
Bishop (2009) obtained some empirical support for her position,
and it seems of relevance in various situations. For example, consider everyday slips and errors mostly reecting inadequate attentional control as assessed by the Cognitive Failures Questionnaire
(CFQ; Broadbent, Cooper, Fitzgerald, & Parkes, 1982). Broadbent
et al. reported a correlation of +.31 between trait anxiety and
CFQ, and Righi et al. (2009) obtained a correlation of +.72. It seems
more plausible that high-anxious individuals experience signicantly more cognitive failures than low-anxious ones on a daily basis because they fail to make use of attentional control mechanisms
rather than because they use them inefciently.
According to attentional control theory (Eysenck et al., 2007),
[Anxiety] is associated with an increased inuence of the stimulus-driven attentional system and a decreased inuence of the
goal-directed attentional system (p. 338). However, the adverse
effects of increased inuence of the stimulus-driven attentional
system can be reduced or eliminated by using compensatory
strategies such as enhanced effort and use of processing resources
(p. 340). Thus, attentional control theory envisages a two-stage
process, although this is not made very explicit in Eysenck et al.
More specically, when the task is undemanding or there is no
clear task goal (e.g., everyday cognitive failures), high-anxious
individuals have a low level of motivation and make minimal use
of attentional control mechanisms. In contrast, when the task is
demanding and there are clear task goals, high-anxious individuals
have a high level of motivation. They attempt to override the inuence of the stimulus-driven attentional system by making extensive use of compensatory strategies (e.g., effortful processing; use
of attentional control mechanisms) to achieve those task goals.

959

Indirect support for the above two-stage process was reported

by Hayes, MacLeod, and Hammond (2009). They used a category
learning task under low and high motivational conditions. High
trait anxiety had a negative effect on performance in the low-motivation condition, but this was eliminated in the high-motivation
condition. These ndings suggest that the increase in effort between the low- and high-motivation conditions was greater for
the high-anxious participants.
In sum, there are two possible effects of anxiety on attentional
control. Anxiety can be associated with decient recruitment of
attentional control resources or it can be associated with substantial (but inefcient) recruitment of such resources. The former is
more likely in conditions involving low motivation (e.g., undemanding task; lack of task goals), whereas the latter is more likely
in conditions involving high motivation (e.g., demanding task;
clear task goals). It is a matter for future research to elucidate more
clearly the circumstances in which each effect is obtained.
References
Ansari, T. L., & Derakshan, N. (in press). Anxiety impairs inhibitory control but not
volitional action control. Cognition & Emotion.
Ansari, T. L., Derakshan, N., & Richards, A. (2008). Effects of anxiety on task
switching: Evidence from the mixed saccade task. Cognitive Affective &
Behavioral Neuroscience, 8, 229238.
Baddeley, A. D. (1986). Working memory. Oxford: Clarendon Press.
Bishop, S. J. (2009). Trait anxiety and impoverished prefrontal control of attention.
Nature Neuroscience, 12, 9298.
Broadbent, D. E., Cooper, P. F., Fitzgerald, P., & Parkes, K. R. (1982). The Cognitive
Failures Questionnaire (CFQ) and its correlates. British Journal of Clinical
Psychology, 21, 116.
Caselli, R. J., Reiman, E. M., Hentz, J. G., Osbourne, D., & Alexander, G. E. (2004). A
distinctive interaction between chronic anxiety and problem solving in
asymptomatic APOE e4 homozygotes. Journal of Neuropsychiatry and Clinical
Neurosciences, 16, 320329.
Christopher, G., & MacDonald, J. (2005). The impact of clinical depression on
working memory. Cognitive Neuropsychiatry, 10, 379399.
Derakshan, N., Ansari, T. L., Hansard, M., Shoker, L., & Eysenck, M. W. (2009).
Anxiety, inhibition, efciency, and effectiveness: An investigation using the
antisaccade task. Experimental Psychology, 56, 4855.
Derakshan, N., & Eysenck, M. W. (2009). Anxiety, processing efciency, and
cognitive performance: New developments from attentional control theory.
European Psychologist, 14, 168176.
Derakshan, N., Smyth, S., & Eysenck, M. W. (2009). Effects of state anxiety on
performance using a task-switching paradigm: An investigation of attentional
control theory. Psychonomic Bulletin & Review, 16, 11121117.
Everling, S., Matthews, A., & Flohr, H. (2001). Prestimulus cortical potentials predict
the performance in a saccadic distractor paradigm. Clinical Neurophysiology, 112,
10881095.
Eysenck, M. W. (1979). Anxiety, learning and memory: A reconceptualization.
Journal of Research in Personality, 13, 363385.
Eysenck, M. W., & Calvo, M. G. (1992). Anxiety and performance: The processing
efciency theory. Cognition & Emotion, 6, 409434.
Eysenck, M. W., Derakshan, N., Santos, R., & Calvo, M. G. (2007). Anxiety and
cognitive performance: Attentional control theory. Emotion, 7, 336353.
Eysenck, M. W., Payne, S., & Derakshan, N. (2005). Trait anxiety, visuospatial
processing, and working memory. Cognition & Emotion, 19, 12141228.
Fales, C. L., Barch, D. M., Burgess, G. C., Schaefer, A., Mennin, D. S., Gray, J. R., et al.
(2008). Anxiety and cognitive efciency: Differential modulation of transient
and sustained neural activity during a working memory task. Cognitive, Affective
& Behavioral Neuroscience, 8, 239253.
Ford, K. A., Goltz, H. C., Brown, M. R. G., & Everling, S. (2005). Neural processes
associated with antisaccade task performance investigated with event-related
fMRI. Journal of Neurophysiology, 94, 429440.
Friedman, N. P., & Miyake, A. (2004). The relations among inhibition and
interference control functions: A latent-variable analysis. Journal of
Experimental Psychology: General, 133, 101135.
Garner, M., Ainsworth, B., Gould, H., Gardner, H., & Baldwin, D. S. (2009). Impaired
attentional control in high and low anxious healthy volunteers: Evidence from
the antisaccade task. European Neuropsychopharmacology, 19. S599-S599.
Goodwin, A. H., & Sher, K. J. (1992). Decits in set-shifting ability in non-clinical
compulsive checkers. Journal of Psychopathology and Behavioral Assessment, 14,
8192.
Hayes, S., MacLeod, C., & Hammond, G. (2009). Anxiety-linked task performance:
Dissociating the inuence of restricted working memory capacity and increased
investment of effort. Cognition & Emotion, 23, 753781.
Hembree, R. (1988). Correlates, causes, effects, and treatment of test anxiety. Review
of Educational Research, 58, 4777.
Jennings, J. R., & van der Molen, M. W. (2005). Preparation for speeded action as a
psychophysiological concept. Psychological Bulletin, 131, 434459.

960

M.W. Eysenck, N. Derakshan / Personality and Individual Differences 50 (2011) 955960

Johnson, D. R. (2009). Emotional attention set-shifting and its relationship to

anxiety and emotion regulation. Emotion, 9, 681690.
Kofman, O., Meiran, N., Greenberg, E., Balas, M., & Cohen, H. (2006). Enhanced
performance on executive functions associated with examination stress:
Evidence from task-switching and Stroop paradigms. Cognition & Emotion, 20,
577595.
Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Howerter, A., & Wager, T.
(2000). The unity and diversity of executive functions and their contributions to
complex frontal lobe tasks: A latent variable analysis. Cognitive Psychology, 41,
49100.
Monsell, S. (2003). Task switching. Trends in Cognitive Sciences, 7, 134140.
Orem, D. M., Petrac, D. C., & Bedwell, J. S. (2008). Chronic self-perceived stress and
set-shifting performance in undergraduate students. Stress: The International
Journal on the Biology of Stress, 11, 7378.
Pacheco-Ungietti, A. P., Acosta, A., Callejas, A., & Lupianez, J. (2010). Attention and
anxiety: Different attentional functioning under state and trait anxiety.
Psychological Science, 21, 298304.
Pacheco-Ungietti, A. P., Lupianez, J., & Acosta, A. (2009). Attention and anxiety:
Relationship between alertness and cognitive control with trait anxiety.
Psicologica, 30, 125.
Righi, S., Mecacci, L., & Viggiano, M. P. (2009). Anxiety, cognitive self-evaluation and
performance: ERP correlates. Journal of Anxiety Disorders, 23, 11321138.

Rosler, F., Heil, M., & Roder, B. (1997). Slow negative brain potentials as reections
of specic modular resources of cognition. Biological Psychology, 45, 109141.
Santos, R., Wall, M., & Eysenck, M. W. (submitted for publication). Anxiety and
processing efciency: fMRI evidence.
Savostyanov, A. N., Tsai, A. C., Liou, M., Levin, E. A., Lee, J.-D., Yurganov, A. V., et al.
(2009). EEG correlates of trait anxiety in the stop-signal paradigm. Neuroscience
Letters, 449, 112116.
Telzer, E. H., Mogg, K., Bradley, B. P., Mai, X., Ernst, M., Pine, D. S., et al. (2008).
Relationship between trait anxiety, prefrontal cortex, and attention bias to
angry faces in children and adolescents. Biological Psychology, 79, 216222.
Tohill, J. M., & Holyoak, K. J. (2000). The impact of anxiety on analogical reasoning.
Thinking and Reasoning, 6, 2740.
Wager, T. D., Jonides, J., & Reading, S. (2004). Neuroimaging studies of shifting
attention: A meta-analysis. Neuroimage, 22, 16791693.
Walkenhorst, E., & Crowe, S. F. (2009). The effect of state worry and trait anxiety on
working processes in a normal sample. Anxiety, Stress & Coping, 22, 167187.
Wieser, M. J., Pauli, P., & Muhlberger, A. (2009). Probing the attentional control
theory in social anxiety: An emotional saccade task. Cognitive, Affective &
Behavioral Neuroscience, 9, 314322.
Wilson, M. R., Vine, S. J., & Wood, G. (2009). The inuence of anxiety on visual
attentional control in basketball free throw shooting. Journal of Sport & Exercise
Psychology, 31, 152168.