Behaviour Research and Therapy 47 (2009) 231236

Contents lists available at ScienceDirect

Behaviour Research and Therapy

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

The clock as a focus of selective attention in those with primary insomnia:

An experimental study using a modied Posner paradigm
H. Woods a, b, *, L.M. Marchetti a, b, S.M. Biello a, C.A. Espie b
a
b

Department of Psychology, University of Glasgow, Glasgow, UK

Glasgow Sleep Centre, Faculty of Medicine, University of Glasgow, Glasgow, UK

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 8 August 2008
Received in revised form
4 November 2008
Accepted 16 December 2008

Espie and colleagues [(2006). The attentionintentioneffort pathway in the development of psychophysiological insomnia: a theoretical review. Sleep Medicine Reviews, 10, 215245] propose a route into
psychophysiological insomnia along the attentionintentioneffort pathway which focuses on the
inhibition of sleep-wake automaticity. A contributing factor to this is selective attention to sleep
(alongside explicit intention to sleep and effort in the sleep engagement process). Following on from
previous work on selective attention to sleep [Marchetti, L. M., Biello, S. M., Broomeld, N. M., MacMahon, K. M. A., & Espie, C. A. (2006). Who is pre-occupied with sleep?. A comparison of attention bias in
people with psychphysiological insomnia, delayed sleep phase syndrome and good sleepers using the
induced change blindness paradigm. Journal of Sleep Research, 15, 212221; MacMahon, K., Broomeld,
N., Macphee, L., & Espie, C. A. (2006). Attention bias for sleep related stimuli in primary insomnia and
delayed sleep phase syndrome using the dot-probe task. Sleep, 29, 11] and considering the importance of
monitoring both internal and external cues in the maintenance of insomnia, as highlighted in the
cognitive model of insomnia [Harvey, A. G. (2002). A cognitive model of insomnia. Behaviour Research and
Therapy, 40, 869893], a cognitive probe task was employed to investigate further the role of the clock as
a focus of selective attention in those with primary insomnia.
A 2  2 between participants design comparing reaction time of individuals with primary insomnia
(n 22) and normal sleepers (n 22) on a modied Posner paradigm. Responses obtained from
a computer task presenting times which fall within a normal sleep period were analysed.
Individuals with primary insomnia demonstrated delayed disengagement to the clock (F(1,84) 6.9,
p < 0.05) which is taken as further support for previous research demonstrating that individuals with
primary insomnia exhibit an attentional bias to sleep related stimuli.
These results lend support to the attentionintentioneffort model (Espie et al., 2006) and the cognitive
model (Harvey, 2002) both of which recognise the importance of selective attention towards salient
stimuli in the maintenance of insomnia. Possible clinical implications of attentional bias to sleep as
a marker of psychopathology progression and treatment efcacy are discussed.
2008 Elsevier Ltd. All rights reserved.

Keywords:
Attention bias
Insomnia
Cognitive arousal
Monitoring
Posner paradigm

Introduction
Insomnia
Primary insomnia (PI) is reportedly found in 3% of the population in western industrialised countries (Ohayon, 1996, 2002).
According to diagnostic criteria, heightened arousal and learned
sleep preventing associations form the foundations of this disorder,
with patients exhibiting excessive focus upon and anxiety about
* Corresponding author. Department of Psychology, University of Glasgow,
Glasgow, Scotland, UK. Tel.: 44 141 330 4757.
E-mail address: heatherw@psy.gla.ac.uk (H. Woods).
0005-7967/$ see front matter 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.brat.2008.12.009

sleep (American Sleep Disorder Association, 1997 & 2005, DSM-IV;

American Psychiatric Association, 1994). Numerous authors
contend that PI is the result of a number of psychological factors,
such as maladaptive beliefs about sleep or excessive pre-sleep
intrusive thoughts (Harvey, 2002; Espie, 2002; Morin, 1993).
Espie, Broomeld, MacMahon, Macphee, & Taylor (2006)
propose a route into PI along the attentionintentioneffort
pathway which focuses on the inhibition of sleepwake automaticity. In their conceptual paper, the authors propose that inhibition
of sleepwake automaticity can be attributed to three processes;
selectively attending to sleep, explicitly intending to sleep and
introducing effort into the sleep engagement process. This model
has been developed by drawing on parallels in the anxiety disorder,

232

H. Woods et al. / Behaviour Research and Therapy 47 (2009) 231236

alcohol and drug abuse literature as well as recent clinical and

experimental studies on insomnia. The cognitive model of insomnia
(Harvey, 2002) likewise highlights the importance of monitoring
both internal and external cues in the maintenance of insomnia.
Selective attention to sleep
An attentional bias is said to have developed when the attention
system becomes sensitive or selective to a particular theme and
impacts on the individuals cognitions. Espie et al. (2006) suggest
that the attentional systems of individuals with PI are particularly
sensitive to sleep. Several studies have now been carried out
establishing a selective attention to sleep in PI using the ICB icker
paradigm and sleep related objects (Marchetti, Biello, Broomeld,
MacMahon, & Espie, 2006), using the dot probe task and sleep
related words (MacMahon, Broomeld, Macphee, & Espie, 2006)
and in a cancer population who have developed sleep onset difculties (Taylor, Espie, & White, 2003).
Clock monitoring
There is widespread evidence that individuals with insomnia
attribute their difculty with sleep to excessive pre-sleep worry
with PI being 10 times more likely to attribute their sleep disturbance to cognitive arousal compared to somatic arousal (Lichstein &
Rosenthal, 1980). The items implicating cognition as the major cause
of the sleep disturbance (e.g. My mind keeps turning things over)
on the Sleep Disturbance Questionnaire are the most highly rated
(Espie, Brooks, & Lindsay, 1989; Harvey, 2001). Therefore, we nd
ourselves at the juncture of having established that pre-sleep worry
contributes to PI but still unable to identify a trigger for such worry.
From clinical practice it is proposed that PI selectively attend to
and monitor for sleep related cues such as body sensations which
are consistent or inconsistent with falling asleep and the environment for signs of not falling asleep (Harvey, 2002). One of the sleep
related cues associated with the stress of not falling asleep is the
bedside clock. Subjective reports highlight two types of clock
monitoring in PI; how long the individual has been in bed without
sleep and how many hours are left before they have to start their
day (Bearpark, 1994; Harvey, 2002). Following from this, Tang,
Schmidt, & Harvey (2007) investigated the association between
clock monitoring, pre-sleep worry and sleep. Thirty-eight PI were
instructed to either monitor a clock or a digital display unit as they
were trying to get to sleep. The clock monitoring task was rated as
more worry provoking and interfered with sleep more than the
digital display monitoring task. The clock monitoring PI reported
more pre-sleep worry and longer SOL on the experimental night
compared to baseline. Hence, in the clock displaying sleep times,
we have an ecologically valid stimulus which is relevant in both the
perspective of the patient and clinical research but also within
everyday life. Most people would be able to recall at least one
occasion where sleep eluded them while being aware the time to
rise was approaching.
Cognitive probe task
The aim of this study was to investigate further the role of the
clock in the development and maintenance of PI, by bringing
together the work discussed above with selective attention in PI
and the real life experimental and clinical work carried out previously which provided valuable subjective reports of the impact of
the clock and time monitoring on sleep in PI.
Although the paradigms previously used to establish an attentional bias to sleep in PI are recognised as legitimate measures of
attentional bias, it was felt that using a modied Posner paradigm

would enable a purer measurement of engagement and disengagement to particular cues presented to the participants.
Posner has suggested that the attention system comprises
measurable cognitive components (shift, engage, disengage; Posner, 1980), which are subserved by specic, neural sub-systems
(Posner & Petersen, 1990) and which are open to modulation by
negative emotional stimuli (Stormark, Laberg, Bjerland, Nordby, &
Hugdahl 1995). In the original Posner cue-target paradigm,
participants responded to a target appearing in the same (valid) or
opposite (invalid) location as a previously presented cue. Results
indicated faster detection of targets on cued trails, particularly at
short (<200 ms) cue-target intervals. This facilitation effect was
taken as evidence of the time-cost of disengaging attention from
the cue to the target on invalid trials (Posner and Petersen, 1990;
Posner, 1988). Over recent years, researchers have begun to apply
Posners attention model to develop a paradigm which can determine whether threatening stimuli can attract attention, i.e.
modulate the engagement component of covert attention, and/or
hold attention, i.e. modulate the disengage component (Broomeld,
Gumley, & Espie, 2005).
Methods
Aims and hypotheses
The aim of this study was two-fold. First, by presenting a pictorial representation of an ecologically valid stimulus, to obtain
objective evidence of the inuence upon attention of monitoring
the clock when set at sleep times and therefore the suggestion of
the clock and sleep time as a precursor to pre-sleep worry. Second,
to explore systematically, with regard to the components of the
attentional system, whether participants would take longer to
categorise the target on invalid trials with PI having the longest
reaction times due to delayed disengagement from this salient cue
compared to normal sleepers (NS).
Design
A 2  2 between-participant design was employed for this
experiment comparing NS and PI on validly and invalidly cued
trials. All subjects recruited completed the computer task followed
by the questionnaires detailed below. On completion of these,
subjects were preliminarily assigned to either NS or PI group. Thus,
it is important to note that the sleep quality for each participant
was not known to the experimenter until the computer task was
completed as well as any priming effect on the participants being
minimised.
Participants
Participants were recruited through advertising online within
the University of Glasgow, a poster campaign around the university
campus and through advertising on the psychology departments
undergraduate student portal for students to obtain course credits.
Each individual was given an appointment time to come into the
department but was given no further information on the nature of
the experiment, other than that it involved a computer based task.
Sleep quality assessment
The sleep quality assessment had three components:
 an interview structured around the ICSD-2 statement of criteria
for psychophysiological insomnia and the DSM-IV statement
of criteria for primary insomnia;

H. Woods et al. / Behaviour Research and Therapy 47 (2009) 231236

 a self-report component where participants completed the

Reduced Horne and Ostberg MorningnessEveningness Questionnaire (MEQ-RF) and the Pittsburgh Sleep Quality Index
(PSQI);
 actigraphy to conrm or discount information gained from the
Horne and Ostberg questionnaire.
After completion of the computer task, those subjects meeting
PI criteria were given an actiwatch to wear and sleep diary to
complete for 5 days. These measures helped to conrm diagnosis of
PI and to discount any circadian disorder based on data obtained
from the MEQ-RF alongside the PSQI such as delayed sleep phase
syndrome. Normal sleepers were not given actiwatches due to the
conrmatory nature of the actigraphy within this study. Actigraphy
was conducted by the Cambridge Neurotechnology Actiwatch
system. This comprises a wrist worn unit ( model AW-4), roughly
the size of a small wristwatch, and sleep analysis software (Actiwatch Sleep Analysis v 1.06), capable of automatically calculating
sleep parameters. There is a general consensus that wrist actigraphy provides an accurate, objective estimate of circadian sleep
and wake parameters (Tyron, 2004).
Inclusion/exclusion criteria
Participants met combined DSM-IV and ICSD-2 criteria for
primary insomnia as well as scoring >6 on the PSQI. PI exclusion
criteria included active psychological or drug interventions for
sleep problems or when a sleep disorder other than insomnia was
suspected. NS were required to score <5 on the PSQI, report no
problems with their sleep and have no history of sleep problems.
For both groups, scoring above 20 on the Beck Depression Inventory
Short Form (BDI-SF4) resulted in exclusion from analyses. The
Spielberger State and Trait Anxiety Inventory was used to provide
a comparative measurement of anxiety between NS and PI but was
not used for exclusion purposes.
All participants had normal vision or corrected to normal vision.
Final allocation of participants to groups for the purposes of analysis did not take place until all sleep assessment data had been
collated. No formally screened participants were excluded due to
either psychological or drug interventions for sleep or because of
substance misuse. However, eight participants were excluded for
failure to return all questionnaire measures.

233

A total of 44 participants were left that met all relevant criteria

and completed all measures, 22 classied into PI and 22 into NS.
A power calculation carried out prior to the study suggested that 21
participants would be required in each of the groups to detect
statistically signicant differences at a power of 0.8 with an alpha
level set at 0.05.
Apparatus and stimuli
A Tiny Mediabook 380 and the experiment-generation package
SuperLab 2.1 (Cedrus Corporation, San Pedro, CA) were used to
implement the Posner paradigm. The size of the screen was 38 cm
(diagonally). Millisecond timing was recorded through SuperLab
Pro and keyboard input.
The sleep related stimuli presented was a photograph of
a digital clock showing a time which is usually associated with
sleeping in a normal sleep pattern, e.g. 02:00 (see Figs. 1 and 2
for diagrammatical representations of stimuli and paradigm). The
stimulus set consisted of 20 digitised pictures of single stimuli.
Twenty pictures were presented on the clock in the left hand box
and 20 in the right hand box. Target stimuli were either two
horizontal dots (..) with a diameter of 0.3 cm or two vertical dots
(:) with a height of 0.3 cm. Cue and target stimuli were presented
inside two boxes (6 cm high and 10.6 cm wide) and positioned
2.0 cm to the left and to the right of the central xation point
(cross shape). These boxes were continually presented on the
screen. Participants were given four practice trials to ensure they
were comfortable with completing the task.
Procedure
Individual trials consisted of a xation cross, presented for 1 s,
followed by the clock picture displayed for 250 ms. Targets were
then presented in the same or different box (central to where one
of the pictures was positioned) and remained on the screen until
response. If the participant saw a target which was vertical, they
were instructed to press C on the computer keyboard, if the target
was horizontal, then they were instructed to press M. When the
target is presented in the box on the same side as the stimuli, this
is considered a valid trial. However, if the target is presented in
the other box on the other side of the computer screen, this is
considered an invalid trial. Latencies to detect these targets were

Fig. 1. Presentation sequence and times of the modied Posner paradigm. The trial is valid when the target is presented on the same side of the screen as the pictorial cue of the
clock showing a sleep time. The trial is invalid when the target is presented on the opposite side of the screen as the cue (as shown above). A valid trial provides a measure of
attentional engagement while an invalid trial provides a measure of attentional disengagement.

234

H. Woods et al. / Behaviour Research and Therapy 47 (2009) 231236

Fig. 2. Enlarged image of the clock showing a sleep time. This is an example of the stimulus used as a cue in the modied Posner paradigm.

used to index the extent to which the groups show an attentional

bias. Participants were asked to complete the task as quickly as
possible and to remain xated on the central cross at all times
(Figs. 1 and 2).
After completion of the task and the questionnaire pack, the true
purpose of the experiment was explained. Participants who were
thought to meet criteria for PI were then asked to wear an actiwatch and to complete a sleep diary for ve nights following the
experiment.
On return of the actiwatch and sleep diary 1 week later, each
participant was given a copy of the Good Sleep Guide [prepared by
Professor Colin A. Espie for the National Medical Advisory
Committee (1994) and now recommended by the British Sleep
Society]. Participants were also provided with contact information
in case they wanted to be informed of the outcome of the project
when it was completed.

of the RT data. This revealed a signicant result for cue location

(F(1,84) 10.53, p >0.01) as well as a signicant two way
interaction between group and cue location (F(1,84) 8.98,
p >0.01). We did not obtain a signicant result for group
(F(1,84) 0.51, p 0.48).
In order to break down the interaction we then examined the
data for sleep quality (NS vs. PI) and cue location (valid vs. invalid)
separately. Scheffe tests were carried out to make comparisons
between means for selected factors. Signicant results were found
between NS and PI for invalid trials (F(1,84) 6.9, p < 0.05) and
between invalid and valid trials for PI (F(1,84) 19.5, p < 0.00001).
These results suggest that PI were slower to respond on invalid
trials than NS suggesting a delayed disengagement from the clock
cue. There was no signicant difference between valid and invalid
trials for NS or between NS or PI on valid trials.
Discussion

Results
The experimental population was equally split between males
and females with a mean age of 24 years. Table 1 shows the
demographics for each sleep quality group (PI and NS) along with
clinical data.
BDI and trait anxiety scores were not signicantly different but
PI participants had higher state anxiety (F(1,42) 4.92, p < 0.05)
which could viably reect an increase in arousal due to presentation of a negative cue.
Table 2 presents summary scores for self-reported sleep data.
Analyses revealed a signicant effect of group at the level of PSQI
with PI having poorer sleep (F(1,42) 160.2, p < 0.0001). Total
sleep time (TST) was also signicantly different (F(1,42) 22.34,
p 0.0001) with PI reporting around 5 h of sleep compared to NS
7 h. Sleep onset latency (SOL) was also signicantly different
(F(1,42 165.88, p 0.0001) with PI taking approximately 45 min
to fall asleep compared to 9 min for NS.
Fig. 3 shows mean (SE) reaction times on both valid and invalid
trials for PI and NS. To assess whether PI was associated with an
attentional bias towards sleep related times, we conducted a 2
(group: PI and NS)  2 (cue location: valid and invalid trial) ANOVA
Table 1
Demographic and clinical data.
PI (n 22)

Age (years)
Gender (% female)
BDI
STAI
SSAI

NS (n 22)

SE

SE

24.4
50
2.1
47.9
38.2

2.1

23.7
50
2.2
44.6
32.2

1.8

0.47
2.48
1.8

0.4
2.6
2.0

Between group
analyses
NS
NS
NS
NS
p < 0.05

BDI, Beck depression inventory; STAI, Spielberger Trait Anxiety Inventory; SSAI,
Spielberger State Anxiety Inventory. Signicant differences were found between NS
and PI on the state anxiety measure which suggests an increase in arousal due to
presentation of a salient cue. NS and PI did not differ in mean age, gender split,
depression or trait anxiety scores.

The purpose of this study was two-fold. First, by presenting

a pictorial representation of an ecologically valid stimulus, our aim
was to obtain objective evidence of the inuence upon attention of
monitoring the clock, when set at sleep related time, and therefore
the suggestion of the clock as a precursor to pre-sleep worry.
Second, we aimed to explore systematically, with regard to the
components of the attentional system, whether participants would
take longer to categorise the target on invalid trials, with PI having
the longest reaction times due to delayed disengagement from this
salient cue compared to NS.
In line with our hypothesis we found PI showed delayed
disengagement from sleep related times. With reference to Fig. 3,
we can see that PI show a speeding effect on valid trials and
a delayed response on invalid trials. The salience of the alarm clock
cue is holding the attention of PI which enables them to process the
target faster when it appears in the position vacated by the cue. This
results in a delaying effect on invalid trials as it slows the movement of attention towards the target on the opposite side of the
screen.
The lack of effect that the sleep related time cue had on NS
reaction time should be considered. Fox, Russo, & Dutton (2002)
used the Posner paradigm to compare high and low trait anxious
people when presented with angry, happy or neutral faces. With
the low anxious group, there was still a main effect for cue validity,
Table 2
Sleep summary data.
PI (n 22)

PSQI
Diary TST (h/min)
Diary SOL (min)

GS (n 22)

SE

SE

10.8
5.4
44.8

0.5
0.2
2.6

3.0
7.2
9.3

0.4
0.3
0.9

Between group
analyses
p < 0.0001
p < 0.0001
p < 0.0001

PSQI, Pittsburgh Sleep Quality Index; TST, total sleep time; SOL, sleep onset latency.
Signicant differences were found between NS and PI on all sleep measures.
Alongside a higher score on the PSQI indicative of poorer sleep, PI took longer to fall
asleep (longer SOL) as well as having less total sleep overall (lower TST).

H. Woods et al. / Behaviour Research and Therapy 47 (2009) 231236

Fig. 3. Reaction time (mean  SE) on valid and invalid trials of NS and PI. Signicant
differences were found between PI and NS on invalid trials and between valid and
invalid trials for PI.

i.e. whether the target was validly or invalidly cued. This was as
expected with longer reaction times for invalid trials compared to
valid trials. This would appear to show the normal time course of
attention to move from one location to another. However, in the
present study, we do not see any effect of cue location in the NS. In
the Fox et al. study, the valence of the face presented as a cue will
still be salient to an extent independent of the anxiety level of the
individual. However, sleep related times may not have an impact on
NS and therefore do not affect the movement of their attention in
anyway. All participants were instructed to keep their eyes xated
on the cross in the centre of the computer screen. This suggests the
sleep time cue was either not registering or was having very little
impact on the attention of normal sleepers which is in direct
contrast to PI who were very much affected by the sleep time cue.
Here we present stimuli which are causing a very different reaction
in individuals with a particular psychopathology compared to those
without, who are displaying no effect at all.
By conrming our hypothesis that PI show a delayed disengagement from sleep related times presented on a clock, we have
provided evidence for an attentional bias towards sleep related
times. These results support previous work carried out using other
paradigms to demonstrate an attentional bias towards sleep related
stimuli (Jones, Macphee, Broomeld, Jones, & Espie, 2005;
MacMahon, Broomeld, Macphee, & Espie, 2006; Marchetti et al.,
2006). However, this study is novel in two ways. First, by focusing
on sleep related times the present study provides objective
evidence for a cognitive process which is argued to be fundamental
in the maintenance of insomnia (Tang et al., 2007).
Second, we used a paradigm which directly measures attention
to a familiar yet clinically relevant stimulus. By selecting a modied
Posner paradigm, we obtain a measure of time taken to disengage
from the clock which can therefore be attributed to attention rather
than response preparation as the location of the cue is not associated with the correct response. Also, the information required to
make the correct response can only be obtained from the cue itself.
This study moves insomnia research forward by objectively
validating the clock as a salient sleep stimulus which impacts on
attentional processing in PI and therefore impacting on sleep. By
establishing an attentional bias to sleep we are able to investigate
further the validity of attentional bias as a marker of disorder
severity and conversely as a marker of treatment efcacy. Other
areas of research such as sleep in cancer patients (Espie et al., 2008)
and smoking abstinence programs (Waters, Shiffman, Bradley, &
Mogg, 2003) have shown attentional bias towards the relevant
salient stimuli to change in accordance with reported treatment

235

success. Bearre, Sturt, Bruce, and Jones (2007) report a relational

attentional bias in clients engaging in a heroin harm reduction
service, i.e. attentional bias towards heroin increases as frequency
of use increases which suggests attentional bias as a potential
indicator of disorder progression.
However, there are methodological limitations to this study.
First, because this was a recruited population, our ability to
generalise to the clinical population may be limited. Attention bias
studies, therefore, require replication with clinical samples.
Nevertheless, it should be noted that non-clinical populations do
display cognitive over-activity at bedtime, with clock monitoring
among the most common (Harvey, 2002). Processes detected at any
stage of sleep disruption, particularly before the clinical extreme,
may provide crucial evidence about the mechanisms involved in
the maintenance/enhancement of the disorder proper.
In conclusion, we used a cognitive probe paradigm to provide
objective evidence for sleep related attentional bias in PI. This
provides support for the previous research done in this eld (Jones
et al., 2005; MacMahon et al., 2006; Marchetti et al., 2006) and
further supports the role of attention in the attentionintention
effort pathway into PI (Espie et al., 2006) by impeding the
automatic passage to sleep. Also, by using the clock and sleep time,
further objective evidence is provided for the role of clock monitoring which has been implicated as a trigger to the cognitive
arousal widely reported by those with PI. The establishment of an
attentional bias in PI provides valuable insight into the development and maintenance of insomnia, however more research is
needed to understand the role of threat and craving as the
underlying mechanisms for selective attention towards sleep and
attentional bias as an indicator of psychopathology progression.

Acknowledgements
This research was supported by the Carnegie Trust for the
Universities of Scotland and a Glasgow University graduate
scholarship.

References
American Psychiatric Association. (1994). Diagnostic and statistical manual of mental
disorders (DSM-IV) (4th ed.). Washington, DC: APA.
American Sleep Disorder Association. (1997). International classication of sleep
disorders (ICSD-2): Revised diagnostic and coding manual. Rochester, MN: APA.
Bearpark, H. (1994). Overcoming Insomnia (what is causing your sleeplessness and
what you can do about it). Australia: Gore & Osment Publications.
Bearre, L., Sturt, P., Bruce, G., & Jones, B. T. (2007). Heroin-related attentional bias
and monthly frequency of heroin use are positively associated in attenders of
a harm reduction service. Addictive Behaviors, 32(4), 784792.
Broomeld, N. M., Gumley, A. I., & Espie, C. A. (2005). Candidate cognitive processes
in psychophysiological insomnia. Journal of Cognitive Psychotherapy, 19, 517.
Espie, C. A. (2002). Insomnia: conceptual issues in the development, persistence and
treatment of sleep disorder in adults. Annual Review of Psychology, 53, 215243.
Espie, C. A., Brooks, D. N., & Lindsay, W. R. (1989). An evaluation of tailored
psychological treatment for insomnia in terms of statistical and clinical
measures of outcome. Journal of Behavior Therapy and Experimental Psychiatry,
20, 143153.
Espie, C. A., Broomeld, N. M., MacMahon, K. M. A., Macphee, L. M., & Taylor, L. M.
(2006). The attentionintentioneffort pathway in the development of psychophysiological insomnia: a theoretical review. Sleep Medicine Reviews, 10, 215245.
Espie, C. A., Fleming, L., Cassidy, J., Samuel, L., Taylor, L. M., White, C. A.,
Douglas, N. J., Engleman, H. M., Kelly, H.-L., & Paul, J. (2008). Randomized
controlled clinical effectiveness trial of cognitive behavior therapy compared
with treatment as usual for persistent insomnia in patients with cancer. Journal
of Clinical Oncology, 26(28), 46514658.
Fox, E., Russo, R., & Dutton, K. (2002). Attentional bias for threat: evidence for
delayed disengagement from emotional faces. Cognition and Emotion, 16(3),
355379.
Harvey, A. G. (2001). Insomnia: symptom or diagnosis? Clinical Psychology Review,
21(7), 10371059.
Harvey, A. G. (2002). A cognitive model of insomnia. Behaviour Research and
Therapy, 40, 869893.

236

H. Woods et al. / Behaviour Research and Therapy 47 (2009) 231236

Jones, B. T., Macphee, L. M., Broomeld, N. M., Jones, B. C., & Espie, C. A. (2005). Sleep
related attentional bias in good, moderate and poor (primary insomnia)
sleepers. Journal of Abnormal Psychology, 114(2), 249258.
Lichstein, K. L., & Rosenthal, T. L. (1980). Insomniacs perceptions of cognitive versus
somatic determinants of sleep disturbance. Journal of Abnormal Psychology, 89,
105107.
MacMahon, K., Broomeld, N., Macphee, L., & Espie, C. A. (2006). Attention bias for
sleep related stimuli in primary insomnia and delayed sleep phase syndrome
using the dot-probe task. Sleep, 29, 11.
Marchetti, L. M., Biello, S. M., Broomeld, N. M., MacMahon, K. M. A., & Espie, C. A.
(2006). Who is pre-occupied with sleep? A comparison of attention bias in
people with psychphysiological insomnia, delayed sleep phase syndrome and
good sleepers using the induced change blindness paradigm. Journal of Sleep
Research, 15, 212221.
Morin, C. M. (1993). Insomnia: psychological assessment and management. New York:
The Guilford Press.
Ohayon, M. (1996). Epidemiological study on insomnia in the general population.
Sleep, 19, S7S15.
Ohayon, M. M. (2002). Epidemiology of insomnia: what we know and what we still
need to learn. Sleep Medicine Reviews, 6, 97111.

Posner, M. I. (1980). Orienting of attention. Quarterly Journal of Experimental

Psychology, 32A, 325.
Posner, M. I. (1988). Structures and functions of selective attention. In T. Boll, &
B. Bryant (Eds.), Master lectures in clinical neuropsychology (pp. 173202).
Washington, DC: Am. Psych. Assoc.
Posner, M. I., & Petersen, S. E. (1990). The attention system of the human brain.
Annual Review of Neuroscience, 13, 2542.
Stormark, K. M., Laberg, J. C., Bjerland, T., Nordby, H., & Hugdahl, K. (1995). Autonomic cued reactivity in alcoholics: the effect of olfactory stimuli. Addictive
Behaviors, 20, 571584.
Tang, N. K. Y., Schmidt, D. A., & Harvey, A. G. (2007). Sleeping with the enemy: clock
monitoring in the maintenance of insomnia. Journal of Behavior Therapy and
Experimental Psychiatry, 38, 4055.
Taylor, L., Espie, C. A., & White, C. A. (2003). Attentional bias in people with acute
versus persistent insomnia secondary to cancer. Behavioral Sleep Medicine, 1(4),
200212.
Tyron, W. W. (2004). Issues of validity in actigraphic sleep assessment. Sleep, 27,
158165.
Waters, A. J., Shiffman, S., Bradley, B. P., & Mogg, K. (2003). Attentional shifts to
smoking cues in smokers. Addiction, 98(10), 14091417.