Behaviour Research and Therapy 40 (2002) 85–103

www.elsevier.com/locate/brat

Is pain-related fear a predictor of somatosensory

hypervigilance in chronic low back pain patients?
a,*
Madelon L. Peters , Johan W. S. Vlaeyen a, Annemarie M. W. Kunnen a, b

a
Department of Medical, Clinical and Experimental Psychology, Maastricht University, 6200 MD Maastricht,
The Netherlands
b
Hoensbroeck centrum voor Arbeidsperspectief, P.O. Box 88, 6430 AB Hoensbroek, The Netherlands
Accepted 12 December 2000

Abstract

Pain-related fear has been found to be associated with increased disability and increased pain perception
in patients with chronic low back pain. A possible mechanism by which pain-related fear could lead to
increased pain perception is heightened attention to somatosensory sensations. In the present study, chronic
pain patients reporting either a high or low level of pain related fear and control participants performed an
auditory reaction time task, while occasionally non-painful electrical stimuli — accompanied by threatening
instructions — were given to the arm or back. In the primary task condition, participants had to perform the
auditory task while ignoring the electrical stimuli. Next, the task was presented under dual task conditions in
which participants had to respond both to tones as well as to detection of electrical stimuli. It was hypothes-
ized that for the primary task, high fearful patients would show greater disruption of performance on the
auditory task than low fearful patients and controls when stimuli were presented to the back. For the dual
task, slower reaction times for the auditory task, in combination with faster detection of electrical stimuli
was expected. The hypotheses were not confirmed but patients scoring high on pain-related fear did show
an overall increase in reaction time for all conditions of the primary task, with or without simultaneous
stimulation. Regression analyses demonstrated that high pain-related fear was associated with increased
reaction time to tones both in patients and healthy controls, and that within patients pain-related fear was
a better predictor of reaction time to tones than present pain intensity. The findings may be interpreted as
showing that patients with elevated levels of pain-related fear habitually attend to somatic sensations, giving
less priority to other attention-demanding tasks.  2001 Elsevier Science Ltd. All rights reserved.

Keywords: Pain-related fear; Chronic low back pain; Primary task paradigm; Attention; Somatosensory stimuli; Hyper-
vigilance

* Corresponding author. Tel.: +31-43-3881603; fax: +31-43-3670968.

E-mail address: madelon.peters@dep.unimaas.nl (M.L. Peters).

0005-7967/01/$ - see front matter  2001 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 0 5 - 7 9 6 7 ( 0 1 ) 0 0 0 0 5 - 5
86 M.L. Peters et al. / Behaviour Research and Therapy 40 (2002) 85–103

1. Introduction

Pain-related fear is considered to be an important factor in the maintenance and exacerbation

of chronic low back pain (CLBP) (Asmundson, Norton & Allerdings, 1997; Asmundson, Norton &
Norton, 1999; Lethem, Slade, Troup & Bentley, 1983; Vlaeyen, Kole Snijders, Boeren & van
Eek, 1995; Vlaeyen & Linton, 2000). Previous studies with CLBP patients have demonstrated
that pain-related fear is associated with substantial avoidance of activities, independent of the
actual amount of pain experienced (Asmundson et al., 1999) and also with functional impairments
as assessed with questionnaires (McCracken, Gross, Aikens & Carnrike, 1996; McCracken, Zay-
fert & Gross, 1992; Vlaeyen, Kole Snijders, Rotteveel, Ruesink & Heuts, 1995). In addition,
patients high on pain-related fear report more intense pain than patients low on pain-related fear
(McCracken et al., 1996) and they report a higher frequency of non-specific somatic sensations
(McCracken, Faber & Janeck, 1998). One of the mechanisms through which pain-related fear
could lead to increased symptom reporting is heightened attention to bodily sensations
(Asmundson et al., 1999). Increased somatosensory attention leads to a higher number of reported
non-specific sensations (Ahles, Cassens & Stalling, 1987; Martin, Ahles & Jeffery, 1991; Schmidt,
Wolfs Takens, Oosterlaan & van den Hout, 1994) and to a higher intensity and frequency of
reported pain (Ferguson & Ahles, 1998). Also, both manipulated attention to pain signals in
healthy volunteers (Arntz & de Jong, 1993; Janssen & Arntz, 1996), as well as self-reported
attention to pain in chronic low back pain patients (McCracken, 1997), is associated with increases
in reported pain.
The assumption that pain-related fear leads to increased attention for pain and perhaps other
somatosensory signals is congruent with cognitive models of fear. These models state that an
important function of fear is the early detection of potentially threatening situations (Eysenck,
1997). Thus, fearful people show enhanced sensitivity and attention for threat stimuli that are
specific for their fears (MacLeod & Mathews, 1988; MacLeod, Mathews & Tata, 1986; McNally,
Riemann & Kim, 1990). A large number of studies have indeed demonstrated fear-specific atten-
tional biases (Logan & Goetsch, 1993; Mathews & MacLeod, 1985; McNally, Riemann & Kim,
1990; Mogg, Mathews & Eysenck, 1992).
So far, little research is available to substantiate the postulate that pain-related fear leads to
increased attention to pain and other somatosensory sensations. Using self-report data, Crombez,
Vervaet, Lysens, Baeyens and Eelen (1998) found an association between avoidance, pain-related
fear and attention to back sensations. In an experimental study, Peters, Vlaeyen and van Drunen
(2000) found that pain-related fear in fibromyalgia patients predicted detection latency of weak
and innocuous electrical stimuli, presented to one of four different body locations. The faster
detection in fearful patients was interpreted as hypervigilance for (innocuous) somatosensory sig-
nals, but other explanations (e.g. sensitization) cannot be ruled out. Experimental studies that are
specifically aimed at measuring selective attention for (painful or non-painful) somatosensory
information in patients reporting pain-related fear can further clarify this issue.
One of the first attempts to measure selective attention to pain-related information in pain
patients with elevated levels of fear was performed by Asmundson, Kuperos and Norton (1997).
Their study showed that in a dot–probe task, controls and patients with low levels of fear directed
their attention away from pain-related stimuli, whereas pain patients with elevated fear scores
attended to pain-related stimuli to the same degree as to neutral stimuli. The predicted increased
 M.L. Peters et al. / Behaviour Research and Therapy 40 (2002) 85–103 87

attention to pain-related stimuli was not found. Another technique that has previously been used
to assess attentional processes depends on establishing deterioration of performance on a so-called
primary task, usually some kind of reaction time or vigilance task. The assumption is that
increased attention to somatosensory signals occurs at the expense of the total attentional capacity
left for performing the primary task. Using the primary task paradigm, it has been shown that
both acute (experimental) and chronic pain demand attention in an automatic fashion and indeed
disrupt performance on ongoing tasks (Crombez, Eccleston Baeyens &, 1995; Eelen, 1996; Eccles-
ton, 1994; Eccleston & Crombez, 1999). Germane to the issue of the role of pain-related fear is
the finding that in addition to pain intensity, specific patient characteristics may play a role in
determining the degree of interruption on the primary task. That is, disruption of attentional per-
formance by pain was found to be most pronounced in chronic pain patients with high somatic
awareness (Eccleston, Crombez, Aldrich & Stannard, 1997) and interestingly, also in patients high
on pain-related fear (Crombez, Eccleston, Baeyens, van Houdenhove & van den Broeck, 1999).
The aim of the present study was to test the assumption that pain-related fear may lead to
enhanced attention for pain and other somatosensory signals. Chronic low back pain patients with
either high or low pain-related fear and controls matched for age, gender and education level,
were subjected to a primary task paradigm. Patients were selected on the basis of their scores on
the Tampa Scale of Kinesiophobia (TSK), a questionnaire measuring fear of (re)injury due to
movement. Scores on this scale have been shown to be predictive of self reported pain and dis-
ability and of experimental tests of avoidance (Crombez et al., 1999; Vlaeyen et al., 1995). Three
groups were formed: (1) patients having scores in the top third of a previously established TSK-
distribution, (2) patients with scores in the bottom third of this distribution, and (3) healthy con-
trols.
All participants performed a reaction time task in which they had to respond as fast as possible
to high and low pitched tones (primary task). This task was presented in isolation, and in combi-
nation with weak electrical stimuli on the back (symptom-specific location) or arm (non-specific
location) in counterbalanced order. It was demonstrated previously that painful stimulation dis-
rupts performance on the primary task (Crombez et al., 1996), but also mildly aversive electrical
stimuli may disrupt performance if these are accompanied by threatening instructions and in
people with a tendency to catastrophize (Crombez, Eccleston, Baeyens & Eelen, 1998a,b). In the
present experiment, we used weak non-painful stimulation to the back and arms, preceded by a
threat inducing instruction. In the first sessions of the task, participants performed the auditory
RT task, while being told to ignore the sensations in arm or back. Thereafter, a dual task was
introduced in which both reaction time in the auditory task, as well as detection of electrical
stimuli (in the back only) was assessed.
As vigilance for sensory signals can be expected to be specifically directed at the location of
the primary pain complaint, we predicted that high fearful patients would show a greater disruption
on the primary task than low fearful patients and controls during trials with simultaneous stimuli
on the back but not during trials with simultaneous stimuli on the arm. During the dual task, we
expected worse performance for high fearful patients on the auditory task, but equal or superior
performance on the body scanning task, due to selective attention to the electrical stimuli in
expense of the attention to tones.
88 M.L. Peters et al. / Behaviour Research and Therapy 40 (2002) 85–103

2. Method

2.1. Participants

Back pain patients were recruited through the pain clinic of the University Hospital Maastricht,
The Hoensbroeck Rehabilitation Center and a private back clinic. The inclusion criterion for back
pain patients was non-specific back pain with a duration of at least 3 months. Eighty-eight patients
were sent information about the study together with a registration form, a questionnaire assessing
demographic variables, pain duration, medication use and somatic/psychiatric comorbidity and
the Dutch version of the Tampa Scale of Kinesiophobia (TSK). Fifty-six patients returned the
registration form and questionnaires. Patients reporting psychiatric and/or serious somatic comor-
bidity, as well as patients using neuroleptic or sedative medication were excluded. Final selection
of patients was based on the TSK scores: patients scoring above 42 (high fearful) and patients
scoring below 35 (low fearful) were included. Cut-off points were determined on the basis of a
previous sample of patients (see below). Controls were recruited by means of a newspaper adver-
tisement, using the same exclusion criteria as for patients.
A total of 59 participants were included in the study, 40 CLBP patients and 19 healthy controls.
Four participants (1 high fearful, 1 low fearful and 2 controls) were replaced by others because
they did not feel the electrical stimuli during the primary task and one participant (control) for
whom the manipulation failed. Groups were matched on age, gender and educational level. Parti-
cipants for the high fearful group were selected first and for each participant in this group the best
possible match was sought. All matched subjects were of the same sex, age deviated maximally 4
years and educational level deviated maximally two levels (using a coding system which classifies
education into 11 levels: 0 being no education and 11 being a university degree). Demographic
data for the 54 participants in the analysis are presented in Table 1. Matching was successful and
there were no significant differences in age (F=1.0, p=0.37) and mean educational level (F=2.3,
p=0.11) among the three groups. The two patient groups did not differ on pain duration (t=⫺0.71,
p=0.48). High fearful patients did seem to take more analgesics and psychotropic medication, but
the difference did not reach significance (medication vs. no medication by group; Chi2=1.8,
p=0.32).

Table 1
Demographic data and pain characteristics for the different groups

High fear n=18 Low fear n=18 Control n=18

Gender (M/F) 8/10 8/10 8/10

Mean age 44.6 46.0 49.0
(sd) (9.7) (11.2) (7.6)
Mean education level 5.6 6.6 7.1
(sd) (2.1) (2.3) (2.1)
Mean pain duration 12.9 10.6
(sd) (9.1) (10.8)
Analgesics 10 6
Psychotropics 4 2
 M.L. Peters et al. / Behaviour Research and Therapy 40 (2002) 85–103 89

2.2. Materials

2.2.1. Tones and shocks

High (1000 Hz) and low (250 Hz) pitch tones were generated by the sound-blaster card of a
PC and delivered to the participant by headphone. Participants responded by pressing the correct
button on a two-button (marked “L” and “H”) console, which terminated the tone instantly. Maxi-
mal duration of the tone was 3000 ms.
Electrical stimuli were delivered by means of 8 mm Ag/AgCL electrodes to either the left and
right ventral underarm or the left and right lower back at L5. The skin under the electrodes was
first rubbed with SkinPure (Nihon Kohden) to reduce skin resistance. During the primary task,
stimuli consisted of 20 ms bipolar square wave pulses delivered for a total duration of 2000 ms
rising to and falling from final amplitude in 200 ms. Amplitude of stimulation was individually
calibrated and varied from 0.7 to 2.5 mA on the arm and 0.7 to 3.4 mA on the back. In the dual
task, stimuli consisted of trains of four bipolar 30 ms square-wave pulses delivered at 1 Hz. The
amplitude of stimulation increased from 0 to 1 mA in 12 s. A second two-button console (marked:
“left” and “right”) was available for responding to the stimuli during the dual task. Delivery of
tones and electrical stimuli were timed by Micro Electronic Laboratory (MEL) Professional
software, version 2 (Schneider, 1995).

2.2.2. Self-report instruments

Pain-related fear: The Dutch version of the Tampa Scale of Kinesiophobia (TSK) was used
for selection purposes. The TSK consists of 17 items, scored on a 4-point scale, measuring fear
of (re)injury due to movement (Kori, Miller & Todd, 1990; Dutch version: Vlaeyen et al., 1995).
Scores can range between 17 and 68. Based on a previous sample of 148 patients (Kole-Snijders,
Vlaeyen, Goossens, Rutten-van Moelken, Heuts, van Breukelen & van Eek, 1999), limits were
established to select patients from the top third and bottom third of the distribution. Low fearful
participants were operationalized as patients with a score of 34 or below and high fearful parti-
cipants were operationalized as patients with a score of 43 or above. Control participants filled
out a slightly modified version of the Dutch TSK (mTSK), adapted for use in non-patient samples
(Vlaeyen & Crombez, 1998). Patients also completed the Pain Anxiety Symptoms Scale (PASS;
McCracken, Zayfert & Gross, 1992). The PASS is a 40 item, 6-point scale questionnaire measur-
ing fear and anxiety behaviors related to pain. It consists of 4 subscales: (1) cognitive anxiety
(PASS-C), (2) fearful appraisals of pain (PASS-F), (3) avoidance behavior (PASS-A) and (4)
somatic symptoms of anxiety (PASS-S).
Somatosensory vigilance: Habitual vigilance for bodily sensations was measured by the Body
Vigilance Scale (BVS; Schmidt, Lerew & Trakowski, 1997; Dutch version: Peters & Vlaeyen,
1996). The BVS is a four-item questionnaire measuring vigilance for sensory symptoms on a 10-
point scale. The last item is an average of the awareness scores of 16 non-specific body symptoms
(e.g. palpitations, dizziness, faintness). In patients, also vigilance for pain sensations was measured
using the Pain Vigilance and Awareness Questionnaire (PVAQ; McCracken, 1997; Dutch version:
Crombez & Vlaeyen, 1998). The instrument contains 16 items that are scored on a 6-point scale.
Negative affectivity: The Dutch translation of the Negative Emotionality Subscale (NEM) of the
Multidimensional Personality Questionnaire (MPQ, Tellegen, 1982; Stegen, Neujens, Crombez,
Hermans, Van de Woestijne & Van den Bergh, 1998) was used. The NEM consists of 14 dichot-
90 M.L. Peters et al. / Behaviour Research and Therapy 40 (2002) 85–103

omous items. A high score on this subscale indicates a nervous, apprehensive, irritable and
emotionally labile personality profile.
Catastrophizing: Catastrophizing cognitions about pain were measured in patients only, using
the Pain Catastrophizing Scale (PCS; Sullivan, Bishop & Pivik, 1995; Dutch version: Crombez &
Vlaeyen, 1996). The PCS consist of thirteen statements concerning pain-related cognitions that
are rated on a 5-point scale.
Pain disability: Patients also completed the Dutch version of the Roland Disability Question-
naire (RDQ; Roland & Morris, 1983; Dutch version: Beurskens, de Vet & Koke, 1996). The RDQ
is a 24-item 2-point scale measuring the extend in which performance of daily activities is ham-
pered due to back pain.
Perceived intensity of stimulation: Rating scales to assess perceived intensity of the electrical
sensations were given to participants after each trial. Participants answered the question “how
intense was the sensation in your right/left arm?” and “how intense was the sensation in your
right/left back?” on a 5 point scale (not felt, just noticeable, clearly noticeable, painful, very
painful).
Pain intensity: patients scored present (back)pain before and after each trial, using a 100 mm
VAS scale anchored no-pain and extreme pain.
Cognitive processes and credibility: At the conclusion of the experimental session, a 10-item
questionnaire was administered to assess self-reported cognitive processes (worrying, distraction)
during the tasks. This questionnaire also included a check for the credibility of the experimental
manipulation (“muscle tension was measured”). This question was imbedded between two neutral
questions about the experimental procedure (e.g. “More high tones were presented than low
tones”) to disguise its meaning and not to alert the participant to the deception. All questions
were answered on a 100 mm VAS-scale, anchored “totally agree” and “do not agree at all”.

2.3. Procedure

2.3.1. Selection and preparation phase

The experiment was approved by the Medical Ethics Committee of the University Hospital
Maastricht. After arrival at the laboratory, participants signed informed consent and completed
some additional questionnaires. Patients also rated present pain intensity (VAS). As part of the
experimental manipulation, participants were given a fictitious explanation of the aim of the
experiment. They were told that muscle tone and back strain due to task performance was assessed.
Using an anatomical model of the spinal column, the experimenter — who wore a nurse’s white
uniform — clarified that strain to the back due to vertical pressure on the spinal column, rotation
or sliding of disks, is often compensated for by increased spinal muscle tone. It was further
explained that increased muscle tone in the lumbar region can be regarded as a signal of back
strain, and that they could feel this as a tingling sensation produced by the electrodes in a feedback
fashion. In reality, weak electrical stimuli controlled by the computer were externally applied. In
order to increase the credibility of the manipulation, it was specified that they would only
occasionally feel the signal, if muscle tone exceeded a pre-set limit a few times. All participants
had to stand upright during the experiment and hold a bag of 1 kg in their non-dominant hand,
and they were told that this could cause strain in the spinal column and joints or tendons of the
 M.L. Peters et al. / Behaviour Research and Therapy 40 (2002) 85–103 91

arm. This procedure was intended to focus participants’ attention on the sensation as a reflection
of an internal process and not as an externally applied stimulus.
Because of different sensitivity across participants, the intensity of the electrical stimulus was
individually calibrated. Care was taken to establish an intensity that was perceivable but not
painful. The rationale given to the participants for the calibration procedure was that the sensitivity
of the apparatus had to be adjusted in order to obtain a reliable signal without the feedback being
painful. The intensity of electrical stimuli was increased in small steps, and participants indicated
when they felt the “feedback” signal. If the participant indicated to have felt something, it was
checked whether it was non-painful. This was repeated for each of the four electrode positions.
If a different stimulus intensity was established for stimuli on the right versus left side of the
arm, the highest intensity was chosen to be used in the experimental trial. It was checked again
whether this intensity was not experienced as painful. The same procedure was used if there was
a difference in intensity for stimuli on the back.

2.3.2. Primary task

The primary task (auditory discrimination) was presented in three trials, each trial lasting
approximately two minutes. One trial consisted of the primary task without distracting stimuli
(non-distraction trial). The other two trials combined the primary task with either distractors on
the arm (arm-distraction trial) or distractors on the lower back (back-distraction trial). The
sequence of trials was counterbalanced across participants so that 3 people in each group received
one of six possible orderings. Each trial consisted of 18 high and 18 low tones. In both distraction
trials, 16 tones with a distractor (equally divided between the high and low tones) and 20 tones
without a distractor were presented in random order. In the back-distraction trial, an equal number
of distractors were presented on the right and the left side. As most strain would be expected to
occur in the arm holding the bag, in the arm-distraction trial 10 stimuli were presented on the
non-dominant arm and 6 on the dominant arm. Inter-stimulus-interval (ISI) between tones varied
between 2000–3000 ms (mean: 2500). During distraction trials, tones were presented 300 ms after
the onset of the distractor stimulus, as a previous study has indicated that the disruptive effect of
pain is most pronounced when tones are presented shortly after pain onset (Crombez et al., 1996).
Tones with and without distractors were randomly distributed across the trials. Reaction time
(RT) to each tone and response accuracy were recorded by MEL.
Before the actual task, participants practiced the auditory discrimination task for 1 minute. It
appeared that all participants understood the task and were able to discriminate the two tones.
Before the back-distraction trial, participants were told that back muscle tone would be measured
during the next two minutes, and that they could occasionally feel a tingling sensation if muscle
tone exceeded a certain limit. It was stressed that they should ignore this sensation and concentrate
on the task. To increase the threat value of the sensations, it was repeated that standing and lifting
could induce strain on the spinal column. Before the arm-distraction trial, similar instructions
were given, but now focused on strain in the arm. After each trial, participants sat down for a
few minutes and completed the rating scale about the intensity of perceived (electrical) sensations,
as well as the VAS present pain (patients only).
92 M.L. Peters et al. / Behaviour Research and Therapy 40 (2002) 85–103

2.3.3. Dual task

After removal of electrodes from the arm, the body scanning task was introduced. In addition
to the tone reaction task, participants were now asked to also take notice of the “feedback” sen-
sation and to respond to it as fast as possible by using a second two-button console. By pressing
the correct button, participants could indicate whether they felt a sensation on the left or right
side. They were instructed that when a sensation was perceived simultaneously with a tone, pri-
ority should be given to the response to the sensation. The threat instruction was repeated and
participants were reminded to pay attention to both the sensations in the back and the auditory task.
Two trials, each consisting of 42 tones (21 high/21 low) and 8 electrical stimuli increasing in
strength, were presented. ISI between the electrical stimuli depended on the time until detection
of the stimulus, but was minimally 8 seconds for undetected stimuli. Each trial lasted approxi-
mately two minutes. Time until stimulus detection, RT for the auditory task and response accuracy
for both tasks were recorded using MEL. During task performance participants remained standing
and holding the bag. After the task, electrodes on the back were removed and patients again rated
their present pain.

2.3.4. Debriefing
After an additional brief computer test (results not reported), the last VAS present pain intensity
was completed. During the subsequent debriefing, participants were explicitly asked whether they
believed that muscle tone had been measured. One participant indicated — at this time, and also
on the manipulation check in the questionnaire — that she had not believed the rationale of the
study, and she was subsequently replaced by another female participant of similar age and edu-
cation. At the conclusion of the session, it was ascertained that each participant understood that
no actual EMG measurement had taken place. A printed version of the debriefing information
was also given to take home. The total procedure lasted about 1.5 hours. An overview of important
parts of the experimental procedure is provided in Table 2.

2.4. Data reduction and analyses

For the primary task, 5 outcome parameters were calculated: (1) RT no distraction (average of
36 tones in the non-distraction trial); (2) RT arm — no distraction (average of 20 tones with no
distractors on the arm, during arm-distraction trial); (3) RT arm — distraction (average of 16
tones with distractors on the arm, during arm-distraction trial); (4) RT back — no distraction
(average of 20 tones with no distractors on the back, during back-distraction trial); (5) RT back —
distraction (average of 16 tones with distractors on the back, during back-distraction trial). Before
averaging RTs, incorrect responses (3%) were removed. RT to high and low pitched tones were
averaged, even though analysis indicated a significant faster RT for high pitched tones. However,
this did not differ between the groups (group×pitch interaction was not significant). RT during
distractors on the right and left arm and RT during distractors on the right and left back were
also averaged, as there appeared to be no difference due to laterality of distractor.
As the two trials of the dual task were identical, RT of high and low pitch tones of both trials
were averaged. Incorrect responses and the response immediately after detection of a stimulus
 M.L. Peters et al. / Behaviour Research and Therapy 40 (2002) 85–103 93

Table 2
Experimental procedure and dependent variablesa

Experimental phase Stimulus set Primary dependent variable

Instruction and
calibration

practice

primary task
counterbalanced

no distraction 36 tones without ES RT no distraction

arm distraction 20 tones without ES RT-arm no distractionb

16 tones with ES on arm RT-arm distractionb

back distraction 20 tones without ES RT-back no distractionb

16 tones with ES on back RT-back distractionb

dual task
trial 1 42 tones
8 ES on the back RT tones dual task
DL stimuli right
trial 2
42 tones DL stimuli left
8 ES on the back
Debriefing

a
ES=electrical stimuli.
b
A secondary variable “delay in RT” was created by subtracting “RT no distraction” from the four other RT variables
in the primary task.

were first removed.1 Also detection latencies (DL) for electrical stimuli were averaged for the
two trials, separately for the 8 stimuli on the right side of the back and for the 8 stimuli on the
left side of the back. Undetected stimuli were assigned the maximal latency time +1 s (1300 ms).
Thus, three parameters were derived from the dual task: (1) RT to tones (overall average of high
and low tones); (2) DL stimuli right back; (3) DL stimuli left back. An overview of the various
dependent variables from each part of the experimental procedure is provided in Table 2 (right
side).
All statistical analyses were performed using SPSS 8.02 for Windows NT. Differences between
the groups on self-report instruments were tested with one-way analysis of variance, or with t-
tests for independent samples when only the two patient groups were considered. The six success-

1
Participants were instructed to give priority to responding to stimulus detection in case a stimulus and tone were perceived
simultaneously. Therefore the response to a tone occurring in proximity of stimulus detection may have been artificially delayed.
94 M.L. Peters et al. / Behaviour Research and Therapy 40 (2002) 85–103

ive pain reports were tested with analysis of variance for repeated measurements with one within-
subject (time) and one between-subject (group) factor.
Data from the primary task were first analyzed by submitting all 5 RT parameters to a multivari-
ate analysis of variance for repeated measures with one within subject factor (measurement) and
one between subject factor (group). Next, the RT from the no-distraction trial was subtracted
from the RTs from the distraction trials, creating four new variables signifying the delay in RT
if distracting stimuli were expected or actually given. A second multivariate analysis of variance
for repeated measurements was carried out on these variables with the within subject factors
“trial” (arm vs. back distraction) and “distraction” (distractor stimuli vs. no distractor stimuli) and
the between subject factor group. The dual task yielded one RT parameter and two DL parameters,
which were subjected to one-way analysis of variance.
Finally, hierarchical regression analyses were carried out to identify predictors for the various
RTs during the primary task. Separate regression models were specified for the total group of
participants and for the two patient groups only. For the total group of participants, predictors
were entered in two steps. In the first step, age and gender were entered as control variables. In
the second step, scores on either the NEM, TSK or BVS were added. Separate analyses were
carried out for each measure, in order to avoid multicollinearity. For patients, predictors were
entered in three steps. The first step consisted of entering the control variables age, gender and
pain duration. In step two, pain was entered, together with the scores on one of the following
questionnaires: NEM, PCS, PVAQ, PASS (subscales and total score) and TSK. Again, successive
analyses were performed for each questionnaire in turn. In step 3 the interaction term between
pain and the questionnaire was entered (e.g. pain intensity×TSK). Before building interaction
terms, questionnaire scores were normalized.

3. Results

3.1. Self report instruments

Scores on the questionnaires are displayed in Table 3. The three groups significantly differed
on the TSK (mTSK for controls), with controls obtaining scores in between the high and low
fearful patient group. Post hoc comparisons indicated that controls significantly differed from high
fearful patients but not from low fearful patients. Differences in BVS scores just failed to reach
significance; NEM scores did not differ between the groups.
High fearful patients scored significantly higher than low fearful patients on the PVAQ, PCS,
RDQ and all subscales of the PASS, except for PASS-C. High fearful patients reported more pain
at baseline, but the difference did not reach significance. Most questionnaires were significantly
intercorrelated (data not shown). The TSK significantly correlated with all other instruments with
the highest correlations obtained for PASS-A (r=0.72), PVAQ (r=0.71) and PCS (r=0.69). TSK
scores did not correlate with baseline pain.
The manipulation check indicated that there were no differences between the groups in credibility
of EMG-measurement (mean VAS scores: 80.2, 87.4 and 91.5, respectively; F(2,51)=1.90,
p=0.161). No differences were found between the groups on the VAS ratings of cognitive pro-
cesses during the task (e.g. distraction, worrying), with one exception. Unexpectedly, high fearful
 M.L. Peters et al. / Behaviour Research and Therapy 40 (2002) 85–103 95

Table 3
Questionnaire scores for the different groupsa

High fear Low fear Control F, tb P

TSK 49.2 (6.43) 28.6 (5.10) 32.8 (6.53) 58.50 ⬍0.001

NEM 4.77 (4.28) 3.22 (3.10) 4.11 (3.92) 0.76 0.476
BVS 17.4 (4.93) 12.7 (6.77) 14.49 (5.35) 3.06 0.056
Pass-C 18.61 (10.83) 12.28 (8.44) 1.96 0.059
Pass-F 14.22 (9.18) 7.28 (4.57) 2.88 0.008
Pass-S 12.11 (10.47) 5.11 (5.63) 2.50 0.019
Pass-A 20.50 (9.33) 11.28 (6.70) 3.41 0.002
Passtot 65.44 (32.41) 35.94 (20.47) 3.27 0.002
PCS 21.50 (9.90) 10.28 (6.09) 4.10 ⬍0.001
PVAQ 45.67 (11.39) 29.39 (8.97) 4.76 ⬍0.001
RDQ 15.11 (5.04) 9.22 (5.53) 3.34 0.002
Present pain 50.10 (26.20) 40.00 (28.70) 1.07 0.295
a
TSK=Tampa Scale of Kinesiophobia; NEM=Negative Emotionality; BVS=Body Vigilance Scale; Pass-C, Pass-F,
Pass-S, Pass-A=Pain Anxiety and Symptom Scale: cognitive, fear, somatic symptoms and avoidance scale; PCS=Pain
Catastrophizing Scale; PVAQ=Pain Vigilance and Awareness Scale; RDQ=Roland Disability Questionnaire.
b
F-statistic is reported when comparing three groups, t-statistic for two groups.

patients scored significantly lower than controls on the question “It was difficult for me to direct
attention to tones due to the muscle tone signals in the arm” (10.5 vs. 31.4; F(2,51)=3.91,
p=0.026). Low fearful patients scored in between the two other groups (16.2).
High fearful patients reported more pain during all 6 measurement periods (50.1, 56.8, 59.1, 60.7,
59.9 and 60.3) than low fearful patients (40.0, 39.9, 41.0, 39.1, 42.1, 41.8) but the group effect
did not reach significance (F(1,32)=3.21, p=0.083). However, there was a significant group×time
interaction (F(5,38)=2.86, p=0.033). Post hoc tests showed that high fearful patients reported more
pain at all 5 time points than low fearful patients, but not during baseline.

3.2. Primary task

First it was tested whether objective or subjective stimulus strength could confound the results.
The three groups did not significantly differ in mean amperage delivered to the arms (1.56, 1.35
and 1.57 mA, respectively, for high fearful, low fearful and control participants; F(2,51)=2.09,
p=0.134) or back (1.96, 1.80 and 1.87 mA; F(2.51)=0.478, p=0.623). Nor were there differences
in perceived sensation between the groups. Mean perceived sensation varied between 1.50 and
1.78 (on a 5-point scale) for all subjects at each of the four positions (all F⬍1 and p⬎0.67)
RT during all three trials (with and without distractors) are displayed in Fig. 1. Multivariate
MANOVA for repeated measurements demonstrated that there was an overall group effect
(F(2,51)=3.25, p=0.047). Post hoc follow-up analyses indicated that high fearful patients differed
significantly from controls on RT in the non-distraction trial, the arm-distraction trial with and
without distractors and the back distraction trial without distractors.
Figure 1 shows that all groups delayed in the back- and arm-distraction trials in comparison
to the non-distraction trial, both with and without the actual presence of distractors. Thus, the
96 M.L. Peters et al. / Behaviour Research and Therapy 40 (2002) 85–103

Fig. 1. Mean reaction time to tones for the three groups of participants during the non-distraction, arm distraction, back
distraction trials and the dual task. Non-distraction, arm distraction and back distraction conditions were administered in
counter-balanced order, the dual task condition was always the last.

mere expectation of muscle tone “feedback” was sufficient to delay RT. The group main effect
for delay in RT did not reach significance (F(2,51)=0.223, p=0.801), indicating that overall, all
groups delayed to the same extent. The trial main effect was not significant (F(1,51)=0.003,
p=0.955), thus average delay was similar in both the arm and back distraction trial. The distractor
main effect did reach significance (F(1,51)=4.57, p=0.037): presence of a distractor stimulus
delayed RT further in comparison to absence of a distractor in the same trial; this was apparent
both in the arm as well as the back distraction condition (distraction×trial effect: F(1,51)=0.018,
p=0.894). The distraction×group interaction did not reach significance ((F(2,51=2.19, p=0.122),
but there was a significant distraction×trial×group interaction (F(2,51)=3.42, p=0.040). However,
in contrast to our prediction, it was the control group that delayed the most when given a dis-
tracting stimulus on the back, whereas the low fearful patients delayed most when given a dis-
tracting stimulus on the arm. High fearful patients were hardly distracted by stimuli on either the
arm or back.

3.3. Dual task

The RT to tones in the dual task can also be seen in Fig. 1. Compared to average RT in all
primary task conditions, RT in the dual task was significantly longer (F(1,51)=69.0, p⬍0.001).
However, there was no group effect on RT in the dual task itself (F(2.51)=2.67, p=0.079), nor
on the delay in RT compared to the primary task condition (F(2,51)=2.98, p=0.60). Detection
latency for stimuli on the left side were 7.4, 6.9 and 6.3 seconds for high fearful, low fearful and
control participants, respectively. For the right side the DLs were: 7.2, 6.2 and 5.2. No significant
 M.L. Peters et al. / Behaviour Research and Therapy 40 (2002) 85–103 97

differences between the group were found in average DL (F(2,51)=0.699, p=0.502) or DL on

either side separately (left side: (F(2,50)=0.85, p=0.434; right side: F(2,50)=2.01, p=0.145).

3.4. Prediction of RT primary task

RTs for the total group of participants from the non-distraction trial, arm-distraction trial (with
and without distractors) and back-distraction trial (with and without distractors) were subjected
to hierarchical regression analyses with age and gender as predictors in step 1 and NEM, BVS
and TSK score as predictors in step 2. Gender was not significantly related to any of the RTs
and subsequently removed in step 1. In step 2, only TSK significantly added to the model. Results
for the five different RTs are shown in Table 4. NEM and BVS did not significantly contribute
to any of the models.
Although results should be interpreted cautiously due to constraints imposed by sample size
(n=36) in combination with multiple analyses, it was explored which variables were related to
RT of patients. Again, age was the only significant predictor in step 1, and gender and pain
duration were removed. In step 2, present pain intensity during that specific trial and — in sub-
sequent analyses — each of the questionnaires were entered. Both the TSK and the PVAQ
appeared to be significant predictors of RT (Table 5). The BVS, NEM, PCS and PASS-subscales
did not contribute to the regression models. None of the interaction terms that were included in
step 3 were significant.

4. Discussion

The present study was designed to test the assumption that patients high on pain-related fear
attend more to somatic sensations, especially sensations related to their pain. The primary outcome
variable was interference on a reaction time task (primary task), due to the administration of non-
painful electrical stimuli to either the arm or back. The results showed that there was an overall
attentional interference effect of stimulation for both the arm and back, indicated by the increased
RTs on the primary task. However, opposite to our prediction, not the high fearful individuals

Table 4
Summary of the hierarchical multiple regression analysis for the total group of participants with mean RT in the various
trials as dependent variablea

No-distraction Arm-distraction⫺ Arm-distraction+ Back-distraction⫺ Back-distraction+

Step 1
Age 0.20 0.29* 0.28* 0.24 0.20
Step 2
TSK 0.39** 0.45** 0.40** 0.39** 0.37
a
Age and gender were entered in step1 and NEM, BVS and TSK in step 2 in separate analyses. Beta-coefficients
for predictors remaining in the final regression model are shown. During the arm and back distraction trials some tones
were presented without simultaneous electrical stimulation (-) and some tones were presented with simultaneous electri-
cal stimulation (+). *=p⬍0.05; **=p⬍0.01.
98 M.L. Peters et al. / Behaviour Research and Therapy 40 (2002) 85–103

Table 5
Summary of the hierarchical multiple regression analysis for the patient group with mean RT in the various trials as
dependent variablea

No-distraction Arm-distraction⫺ Arm-distraction+ Back-distraction⫺ Back-distraction+

Step 1
Age 0.28 0.41* 0.34 0.49* 0.25
For models with the TSK as predictor
Step 2
Pain ⫺0.11 ⫺0.05 ⫺0.07 ⫺0.14 ⫺0.20
TSK 0.38* 0.41* 0.37* 0.34 0.39*
Step 3
TSK×pain ⫺0.15 ⫺0.00 0.01 ⫺0.11 ⫺0.13
For models with the PVAQ as predictor
Step 2
Pain ⫺0.15 ⫺0.11 ⫺0.14 ⫺0.21 ⫺0.28
PVAQ 0.41* 0.50** 0.44* 0.44* 0.52**
Step 3
PVAQ×pain 0.23 ⫺0.08 ⫺0.08 0.24 ⫺0.26
a
Age, gender and pain duration were entered in step 1. In step 2 present pain intensity was entered together with
one of the following variables (in separate analyses): NEM, BVS, PCS, PASS-A, PASS-C, PASS-S, PASS-F, PASS-
total, TSK and PVAQ Beta-coefficients for predictors remaining in the final regression model are shown. *
p⬍0.05; **p⬍0.01.

but the healthy controls showed the highest degree of interference when stimuli were given on
the back, compared to their performance in the same condition without stimuli. Thus, the hypoth-
esis that patients scoring high on pain-related fear are more distracted from the primary task under
conditions of back-stimulation could not be confirmed. In the dual task, no significant differences
were found between the groups, neither in RT to tones nor in detection latency of electrical stimuli.
The observation of an overall distraction effect of stimulation in the primary task agrees with —
and extends — the findings of Crombez et al. (1998b). In that study it was demonstrated that
mildly aversive electrocutaneous stimuli lead to increased RTs in the primary task and,
importantly, if these stimuli were accompanied by threat instructions, the interference effect
increased. In the present study, weak innocuous stimuli were used, and it appeared that these
weak stimuli were also able to induce an attentional interference. We tried to make the sensations
threatening by telling participants that they reflected strain in the joints and tendons. We had
expected that such an explanation would make the stimuli especially threatening for back pain
patients high on pain-related fear. However, all participants showed a delay in RT during stimu-
lation, and if anything, controls seemed to delay most. Whether the threat manipulation actually
worked, and more so for patients with pain-related fear, is uncertain as we did not directly assess
perceived threat.
Another extension of previous findings is the fact that our study shows that the mere expectation
of upcoming signals can induce an attentional shift away from the primary task, as demonstrated
by the substantial increase in RT during the arm and back distraction conditions in the absence
of actual stimulation. This increase in RT in both distraction conditions cannot be attributed
 M.L. Peters et al. / Behaviour Research and Therapy 40 (2002) 85–103 99

to an order effect, as conditions were presented in counterbalanced order. Thus, anticipation of

somatosensory sensations may be enough to interrupt ongoing performance.
Even though pain-related fear did not lead to a greater disruptive effect on the primary task
during stimulation, it is of interest to note that there was an overall difference in RT between the
three groups. Patients with high pain-related fear had the largest RTs during all phases of the
experiment and controls had the smallest RTs. Moreover, RTs in all conditions were predicted
by pain-related fear as measured by the TSK, for all participants as well as within the patient
group, and pain-related fear proved to be a more powerful predictor than reported pain intensity.
Post hoc, this may tentatively be interpreted to reflect the tendency of patients with elevated levels
of pain-related fear to habitually attend to somatic sensations, giving less priority to other atten-
tion-demanding tasks. Thus, their performance on the primary task could have been slowed down
already under baseline conditions. The anticipation of sensations in the body (arm or back), may
have increased vigilance towards the body to the same extent in all three groups, independent of
pain-related fear. The fact that high fearful patients did not show an additional interference effect
if stimuli were actually given, may be interpreted as signifying that these patients were already
attending towards bodily sensation in a large degree.
An alternative explanation for the differences between the groups in overall RT could be the
intensity of pain that patients experienced during the task. High fearful patients reported more
pain during performance of the tasks than low fearful patients. Pain automatically attracts atten-
tion, leading to decreased performance on RT tasks and high intensity pain seems to attract more
attention than low intensity pain (Eccleston 1994, 1995; Eccleston et al., 1997). Although we
cannot entirely rule out this explanation, in the regression analysis, pain-related fear was a better
predictor of RT than present pain intensity. Present pain intensity did not significantly contribute
to any of the models; moreover beta coefficients were negative while a positive association would
have been expected.
A recent study by Crombez et al. (1999) using a primary task paradigm found that neither pain
intensity, nor pain related-fear as measured by the TSK, were predictors of attentional interference.
However, the interaction between pain intensity and pain-related fear significantly predicted per-
formance on the primary task. We could not replicate the interaction effect between pain intensity
and pain-related fear, but instead found a main effect for pain-related fear. One explanation could
be that by our deliberate selection of the one-third highest and one-third lowest scorers on the
TSK, our patients varied more in pain-related fear than the sample in the study of Crombez et
al. In fact, the average TSK score obtained for patients in that study (44.3) indicated a rather
high degree of fear for the total patient sample.
Interestingly, RT to tones in the primary task was also predicted by pain vigilance as measured
by questionnaires (PVAQ). Thus, patients who reported that they are vigilant for pain signals,
performed less well on the auditory task than patients reporting less attention to pain. Thus again,
heightened attention to pain appears to interrupt performance on the auditory task during all phases
of the experiment. However, a methodological concern in the interpretation of the regression
analyses may be that separate analyses were performed for each of the five dependent RT variables
and with various combinations of predictors. Although it should be admitted that this increases
the likelihood of chance findings, the outcomes are consistent across all analyses, i.e. the same
variable (PVAQ, TSK) is predictive of RT at each of the five different time points. Results are
also consistent with the notion that more proximally related factors have a stronger association
100 M.L. Peters et al. / Behaviour Research and Therapy 40 (2002) 85–103

with the dependent variable than distally related factors. According to the previously described
cognitive model of chronic pain, negative affectivity and catastrophizing predisposes towards
pain-related fear, and pain-related fear leads to somatosensory vigilance and — as proposed —
disruption of attention to other tasks. Thus, the fact that the strongest association was found
between pain vigilance and RT, followed by pain-related fear and RT is consistent with their
more proximal position in the theoretical model.
In the dual task, in which participants performed the same RT task with high and low tones,
but now also had to attend to the sensations in their back, no differences were found between
the groups in RT for tones. The differences between the groups in RT were actually smaller —
and non-existent for the two patient groups — in the dual task than during all primary task
conditions. This may be explained by the fact that during the dual task all participants were
instructed to attend to the sensations in the back. In fact, we asked them to give priority to
responding to the sensations instead of responding to tones. Thus, in retrospect, the finding that
controls and low fearful patients showed similar delays in RT during the dual task may not be
surprising: a somatic focus was experimentally induced in all participants, overruling pre-existing
differences between individuals.
In a previous experiment using fibromyalgia patients, we found that patients high on pain-related
fear showed superior detection of weak electrical stimuli on the arm or leg (Peters, Vlaeyen & van
Drunen, 2000). In the present experiment, detection of similar electrical stimuli on the back was
assessed. Superior detection in the dual task could not be replicated. High fearful patients even
seemed to respond somewhat slower, although the difference did not reach significance. Again, this
may be due to the instruction of giving priority to the detection of sensations in the back. In the
previous experiment the focus of attention was left open, and participants could decide for themselves
to which task they gave priority. Also, as the focus of the present study was on the primary task, a
less elaborate method of detection latency measurement was used. Less stimuli than in the previous
experiment were given and stimulus intensity rose much faster, thereby decreasing sensitivity of the
measure. Thus, it remains to be determined whether detection threshold and detection accuracy are
associated with pain-related fear in low back pain patients.
Evidence for the association between pain-related fear and somatic attention was obtained in
this experiment from the self-report measures. Although all self-report measures correlated fairly
high, the highest correlation was obtained between the TSK and pain avoidance as measured by
the PASS and — pertinent to our question — between the TSK and pain vigilance as measured
by the PVAQ. This is in accordance with our previous study with fibromyalgia and with a study
by Crombez et al. (1999), both reporting significant associations between pain-related fear and
attention to pain.
In conclusion, neither the primary task nor the dual task could confirm our a priori prediction
that patients with high pain-related fear would show more interference on the (primary) tone task
in the presence of stimuli on the back than patients with low-pain related fear and controls. Pain-
related fear was associated with an overall deterioration of performance on the primary task, but
whether this is actually due to a habitually increased attention towards the body remains to be
determined. Self-report measures did seem to indicate an association between pain-related fear
and vigilance for pain sensations. Further studies aimed at elucidating the link between pain-
related fear and hypervigilance are being carried out at present, using both the primary task
paradigm as well as more direct tests of attentional bias like stroop and dot–probe procedures.
 M.L. Peters et al. / Behaviour Research and Therapy 40 (2002) 85–103 101

Acknowledgements

We wish to thank Mr G. Westhovens for his contribution to patient selection and Mr Bert
Hoekzema for realizing the software for the experiment. Mrs Suzanne Stomp-van de Berg of the
Pain Management and Research Center, University Hospital Maastricht, is acknowledged for her
help by patient selection. This study was supported by a grant from The Council for Medical and
Health Research of the Netherlands (MW-NWO: 904-65-090).

References

Ahles, T. A., Cassens, H. L., & Stalling, R. B. (1987). Private body consciousness, anxiety and the perception of pain.
Journal of Behavioral Therapy and Experimental Psychiatry, 18, 215–222.
Arntz, A., & de Jong, P. F. (1993). Anxiety, attention and pain. Journal of Psychosomatic Research, 37 (4), 423–431.
Asmundson, G. J., Kuperos, J. L., & Norton, G. R. (1997). Do patients with chronic pain selectively attend to pain-
related information? Preliminary evidence for the mediating role of fear. Pain, 72, 27–32.
Asmundson, G. J., Norton, G. R., & Allerdings, M. D. (1997). Fear and avoidance in dysfunctional chronic back pain
patients. Pain, 69, 231–236.
Asmundson, G. J., Norton, P. J., & Norton, G. R. (1999). Beyond pain: the role of fear and avoidance in chronicity.
Clinical Psychology Review, 19, 97–119.
Asmundson, G. J., & Taylor, S. (1996). Role of anxiety sensitivity in pain-related fear and avoidance. Journal of
Behavioral Medicine, 19, 577–586.
Beurskens, A. J., de Vet, H. C., & Koke, A. J. (1996). Responsiveness of functional status in low back pain: a compari-
son of different instruments. Pain, 65, 71–76.
Crombez, G., Eccleston, C., Baeyens, F., & Eelen, P. (1996). The disruptive nature of pain: An experimental investi-
gation. Behaviour Research and Therapy, 34, 911–918.
Crombez, G., Eccleston, C., Baeyens, F., & Eelen, P. (1998a). Attentional disruption is enhanced by the threat of pain.
Behaviour Research and Therapy, 36 (2), 195–204.
Crombez, G., Eccleston, C., Baeyens, F., & Eelen, P. (1998b). When somatic information threatens, catastrophic think-
ing enhances attentional interference. Pain, 75, 187–198.
Crombez, G., Eccleston, C., Baeyens, F., van Houdenhove, B., & van den Broeck, A. (1999). Attention to chronic
pain is dependent upon pain-related fear. Journal of Psychosomatic Research, 47, 403–410.
Crombez, G., Vervaet, L., Lysens, R., Baeyens, F., & Eelen, P. (1998). Avoidance and confrontation of painful, back-
straining movements in chronic back pain patients. Behavior Modification, 22, 62–77.
Crombez, G., Vlaeyen, J. W., Heuts, P. H., & Lysens, R. (1999). Pain-related fear is more disabling than pain itself:
evidence on the role of pain-related fear in chronic back pain disability. Pain, 80, 329–339.
Crombez, G., & Vlaeyen, J. W. S. (1996). The Pain Catastrophizing Scale: Unpublished authorized Dutch/Flemish trans-
lation.
Crombez, G., & Vlaeyen, J. W. S. (1998). The Pain Vigilance and Awareness Questionnaire. Unpublished authorized
Dutch/Flemish translation.
Eccleston, C. (1994). Chronic pain and attention: A cognitive approach. British Journal of Clinical Psychology, 33,
535–547.
Eccleston, C. (1995). Chronic pain and distraction: an experimental investigation into the role of sustained and shifting
attention in the processing of chronic persistent pain. Behaviour Research and Therapy, 33, 391–405.
Eccleston, C., & Crombez, G. (1999). Pain demands attention: a cognitive–affective model of the interruptive function
of pain. Psychological Bulletin, 125, 356–366.
Eccleston, C., Crombez, G., Aldrich, S., & Stannard, C. (1997). Attention and somatic awareness in chronic pain. Pain,
72, 209–215.
Eysenck, M. W. (1997). Anxiety and cognition: A unified theory. Hove: Psychology Press/Erlbaum.
102 M.L. Peters et al. / Behaviour Research and Therapy 40 (2002) 85–103

Ferguson, R. J., & Ahles, T. A. (1998). Private body consciousness, anxiety and pain symptom reports of chronic pain
patients. Behaviour Research and Therapy, 36, 527–535.
Grisart, J. M., & Plaghki, L. H. (1999). Impaired selective attention in chronic pain patients. European Journal of
Pain, 3, 325–333.
Janssen, S. A., & Arntz, A. (1996). Anxiety and pain: Attentional and endorphinergic influences. Pain, 66, 145–150.
Kole-Snijders, A. M. J., Vlaeyen, J. W. S., Goossens, M. E. J. B., Rutten-van Moelken, M. P. M. H., Heuts, P. H. T.
G., van Breukelen, G., & van Eek, H. (1999). Chronic Low-Back Pain: What does cognitive coping skills training
add to operant behavioral treatment? Results of a randomized clinical trial. Journal of Consulting and Clinical
Psychology, 67, 931–944.
Kori, S. H., Miller, R. P., & Todd, D. D. (1990). Kinesiophobia: A new view of chronic pain behavior. Pain Manage-
ment, Jan/Feb, 35-43.
Lethem, J., Slade, P. D., Troup, J. D., & Bentley, G. (1983). Outline of a fear-avoidance model of exaggerated pain
perception: I. Behaviour Research and Therapy, 21, 401–408.
Logan, A. C., & Goetsch, V. L. (1993). Attention to external threat cues in anxiety states. Clinical Psychology Review,
13, 541–559.
MacLeod, C., & Mathews, A. (1988). Anxiety and the allocation of anxiety to threat. The Quarterly Journal of Experi-
mental Psychology, 40a, 653–670.
MacLeod, C., Mathews, A., & Tata, P. (1986). Attentional bias in emotional disorders. Journal of Abnormal Psychology,
95, 15–20.
Martin, J. B., Ahles, T. A., & Jeffery, R. (1991). The role of private body consciousness and anxiety in the report of
somatic symptoms during magnetic resonance imaging. Journal of Behavior Therapy and Experimental Psychiatry,
22 (1), 3–7.
Mathews, A., & MacLeod, C. (1985). Selective processing of threat cues in anxiety states. Behaviour Research and
Therapy, 23, 563–569.
McCracken, L. M. (1997). “Attention” to pain in persons with chronic pain: A behavioral approach. Behavior Therapy,
28, 271–284.
McCracken, L. M., Faber, S. D., & Janeck, A. S. (1998). Pain-related anxiety predicts non-specific physical complaints
in persons with chronic pain. Behavior Research and Therapy, 36, 621–630.
McCracken, L. M., Gross, R. T., Aikens, J., & Carnrike, C. L. M. Jr. (1996). The assessment of anxiety and fear in
persons with chronic pain: A comparison of instruments. Behaviour Research and Therapy, 34, 927–933.
McCracken, L. M., Zayfert, C., & Gross, R. T. (1992). The Pain Anxiety Symptoms Scale: Development and validation
of a scale to measure fear of pain. Pain, 50, 67–73.
McNally, R. J., Riemann, B. C., & Kim, E. (1990). Selective processing of threat cues in panic disorder. Behaviour
Research and Therapy, 28, 407–412.
McNeil, D. W., & Rainwater, A. J. (1998). Development of the Fear of Pain Questionnaire–III. Journal of Behavioral
Medicine, 21, 389–410.
Mogg, K., Mathews, A., & Eysenck, M. (1992). Attentional bias to threat in clinical anxiety states. Cognition and
Emotion, 6, 149–159.
Peters, M. L. & Vlaeyen, J. W. S. (1996). The Body Vigilance Scale: Unpublished Dutch translation.
Peters, M. L., Vlaeyen, J. W. S., & van Drunen, C. (2000). Do fibromyalgia patients display hypervigilance for innocu-
ous somatosensory stimuli? Pain, 86, 283–292.
Roland, M., & Morris, R. (1983). A study of the natural history of back pain. Part I: development of a reliable and
sensitive measure of disability in low-back pain. Spine, 8, 141–144.
Schmidt, A. J. M., Wolfs Takens, D. J., Oosterlaan, J., & van den Hout, M. A. (1994). Psychological mechanisms in
hypochondriasis: Attention-induced physical symptoms without sensory stimulation. Psychotherapy and Psychosom-
atics, 61, 117–120.
Schmidt, N. B., Lerew, D. R., & Trakowski, J. H. (1997). Body vigilance in panic disorder: evaluating attention to
bodily perturbations. Journal of Consulting and Clinical Psychology, 65, 214–220.
Schneider, W. (1995). MEL professional user’s guide (version 2). Pittsburgh, PA: Psychological Software Tools, Inc.
Stegen, K., Neujens, A., Crombez, G., Hermans, D., Van De Woestijne, K. P., & Van den Bergh, O. (1998). Negative
affect, respiratory reactivity, and somatic complaints in a CO2 enriched air inhalation paradigm. Biological Psy-
chology, 49, 109–122.
 M.L. Peters et al. / Behaviour Research and Therapy 40 (2002) 85–103 103

Sullivan, M. J. L., Bishop, S. R., & Pivik, J. (1995). The Pain Catastrophizing Scale: Development and validation.
Psychological Assessment, 7, 524–532.
Tellegen, A. (1982). Brief Manual for the Differential Personality Questionnaire (Unpublished Report). Minnesota:
University of Minnesota.
Vlaeyen, J. W. S. & Crombez, G. (1998). The modified Tampa Scale of Kinesiophobia: for its use in non-pain samples;
Unpublished Dutch/Flemish translation.
Vlaeyen, J. W., Kole Snijders, A. M., Boeren, R. G., & van Eek, H. (1995). Fear of movement/(re)injury in chronic
low back pain and its relation to behavioral performance. Pain, 62, 363–372.
Vlaeyen, J. W. S., Kole Snijders, A. M. J., Rotteveel, A. M., Ruesink, R., & Heuts, P. (1995). The role of fear of
movement/(re)injury in pain disability. Journal of Occupational Rehabilitation, 5, 235–252.
Vlaeyen, J. W. S., & Linton, S. J. (2000). Fear-avoidance and its consequences in chronic musculoskeletal pain: state
of the art. Pain, 85, 317–332.