Rapid Communication Cerebral activation during hypnotically induced and imagined pain

Abstract

The continuing absence of an identifiable physical cause for disorders such as chronic low back pain, atypical facial pain, or fibromyalgia, is a source of ongoing controversy and frustration among pain physicians and researchers. Aberrant cerebral activity is widely believed to be involved in such disorders, but formal demonstration of the brain independently generating painful experiences is lacking. Here we identify brain areas directly involved in the generation of pain using hypnotic suggestion to create an experience of pain in the absence of any noxious stimulus. In contrast with imagined pain, functional magnetic resonance imaging (fMRI) revealed significant changes during this hypnotically induced (HI) pain experience within the thalamus and anterior cingulate (ACC), insula, prefrontal, and parietal cortices. These findings compare well with the activation patterns during pain from nociceptive sources and provide the first direct experimental evidence in humans linking specific neural activity with the immediate generation of a pain experience.

Full text

Rapid Communication
Cerebral activation during hypnotically induced and imagined pain
Stuart W.G. Derbyshire,
a,
*
Matthew G. Whalley,
b
V. Andrew Stenger,
c
and David A. Oakley
b
a
Departments of Anesthesiology and Radiology, University of Pittsburgh Medical Center, Pittsburgh, PA 15213, USA
b
Department of Psychology, Hypnosis Unit, University College London, London WC1E 6BT, UK
c
Department of Radiology, University of Pittsburgh Medical Center, Pittsburgh, PA 15213, USA
Received 3 December 2003; revised 24 February 2004; accepted 28 April 2004
The continuing absence of an identifiable physical cause for disorders
such as chronic low back pain, atypical facial pain, or fibromyalgia, is a
source of ongoing controversy and frustration among pain physicians
and researchers. Aberrant cerebral activity is widely believed to be
involved in such disorders, but formal demonstration of the brain
independently generating painful experiences is lacking. Here we
identify brain areas directly involved in the generation of pain using
hypnotic suggestion to create an experience of pain in the absence of
any noxious stimulus. In contrast with imagined pain, functional
magnetic resonance imaging (fMRI) revealed significant changes
during this hypnotically induced (HI) pain experience within the
thalamus and anterior cingulate (ACC), insula, prefrontal, and parietal
cortices. These findings compare well with the activation patterns
during pain from nociceptive sources and provide the first direct
experimental evidence in humans linking specific neural activity with
the immediate generation of a pain experience.
D 2004 Elsevier Inc. All rights reserved.
Keywords: Functional magnetic resonance imaging; Hypnotically induced
pain; Physically induced pain; Anterior cingulate cortex
Introduction
An extensive functional imaging literature has demonstrated
that pain experience is mediated via activation of a network of
cortical regions including the anterior cingulate cortex (ACC),
insula, prefrontal regions, and primary (S1) and secondary (S2)
somatos ensory cortices (Derbyshire et al., 2002; Casey, 1999;
Peyron et al., 2000; Price, 2000; Treede et al., 1999).The
interpretation of these findings is complicated, however, by pro-
cesses associated with the stimulus that are incidental to the actual
sensory and emotional experience of pain. Such processes include
motor inhibition or motor control responses and processes attrib-
utable to the innocuous components of the stimulus. A technique
that provides for painful experience in the absence of stimulation
would be valuable in the identification of brain regions that are
critically and uniquely associated with the sensory and emotional
components of pain.
Such a technique would also be valuable in identifying
regions of the brain that may be actively generating pain
disorders in patients where other abnormality cannot be demon-
strated. Abnormal activation within the pain network has been
postulated to cause or partially generate certain clinical pain
disorders such as chronic low back pain, atypical facial pain,
and fibromyalgia (Derbyshire et al., 1994, 2002; Gracely et al.,
2002). Such disorders fa ll broadly under the umb rella of
functional pain, defined as consisting of one or more symptoms
that, after appropriate medical assessment, cannot be explained
in terms of a conventionally defined medical disease (Wessely et
al., 1999). This exclusory definition is clearly problematic
because the possibility of future diagnosis based on objective
findings remains open and unresolved (Derbyshire, 1999). Ele-
vated spinal fluid substance P, abnormal single photon emission
computed tomography (SPECT) and functional magnetic reso-
nance imaging (fMRI) scans, and low serum growth hormone
levels, as described in fibromyalgia patients (Bennett et al.,
1997; Gracely et al., 2002; Mountz et al., 1995; Russell et al.,
1994), might be precursors of a ‘conventional’ medical diagno-
sis. A model of functional pain based upon early or greater
activation of central regions responsible for pain experience
might also be integrated into a biomedical understanding of
functional disorder (Croft, 2000; Derbyshire et al., 1994, 2002;
Gracely et al., 2002).
Nevertheles s, the known intercon nection of stress, negative
affect, and pain has led to suggestions that various stimuli ranging
from injury elsewhere in the body to emotional and cognitive inputs
from higher neural centers can expand, amplify, or create pain
symptoms (Croft, 2000; Derbyshire, 2004). Taken together, these
hypotheses and data raise the possibility that an experience of pain
can originate exclusively within a subject’s brain or mind rather than
being necessarily dependent on the pathology of peripheral tissue.
The existence of a neural functional pain mechanism is
supported by studies that have shown brain activation to be
generally colinear with reported pain experience, rather than
stimulus intensity, and by demonstration of specific modulation
of brain activity via manipulation of affective and sensory
dimensions of pain experience (Coghill et al., 2003; Croft,
2000; Derbyshire et al., 1997, 2002; Faymonville et al., 2003;
1053-8119/$ - see front matter D 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.neuroimage.2004.04.033
* Corresponding author. University of Pittsburgh Medical Center, MR
Research Center, 200 Lothrop Street, Pittsburgh PA 15213. Fax: +1-412-
647-9800.
E-mail address: derbyshiresw@anes.upmc.edu (S.W.G. Derbyshire).
Available online on ScienceDirect (www.sciencedirect.com.)
www.elsevier.com/locate/ynimg
NeuroImage 23 (2004) 392 – 401

Gracely et al., 2002; Rainville et al., 1997). The extent to which
different cortical structures might actively generate a painful
experience independent of peripheral input, however, is largely
unknown and untested.
We have previously argued that although there are differ-
ences, in context and chronicity for example, there are common
mechanisms underlying functional neurological symptoms, such
as those seen in conversion disorder, and in comparable
phenomena produc ed by suggestion in hypn osis (Oakley,
1999). In support of this view, similar patterns of brain
activation have been demonstrated during attempted movement
in a subject with a hypnotically induced (HI) lower limb
paralysis (Halligan et al., 2000) and in a comparable conversion
disorder patient (Marshall et al., 1997). There is also evidence
that hypnotically induced paralysis is not only experienced as an
involuntary effect but is mediated by different brain processes
compared to the mere simulation or imitation of the same
paralysis in hypnotized subjects (Oakley et al., 2003; Ward et
al., 2003). These observations raise the possibility that a similar
commonality in mechanism may exist in clinically encountered
functional pain conditions and in the experience of hypnotically
induced pain.
There is already some evidence that hypnotic suggestion can
be used to produce the experience of pain in the absence of a
physical stimulus with concomitant changes in galvanic skin
response (GSR), heart rate, and respiration (Barber and Hahn,
1964; Dudley et al., 1966; Hilgard et al., 1974). More recently,
we have investigated similarities in participants’ experiences of
hypnotically induced and physically induced (PI) pains (Whalley
and Oakley, 2003), demonstrating the induction of a painful
sensation in the absen ce of a physical stimulus. Functional
imaging techniques offer the opportunity to objectively validate
such self-report measures of pain. Activation of the pain
network in the absence of noxious stimulation would support
the possibility of direct central involvement in functional pain
disorders.
In the present study, we used hypnosis as a cognitive tool to
reveal cerebral mechanisms of pain generation in normal human
volunteers (Rainville et a l., 1 997; Raz and Shapiro, 2002).A
perceptual experience of pain was achieved with a hypnotic induc-
tion followed by the suggestion of painful heat without actual
delivery of any stimulus (Whalley and Oakley, 2003). Cerebral
cortical activity related to this hypnotically induced functional pain
experience was measured using functional magnetic resonance
imaging and compared with activation during actual delivery of
noxious heat.
Materials and methods
Hypnosis
Functional pain was induced using a procedure adapted from
Szechtman et al. (1998) and confirmed in the individuals chosen
for these imaging experiments. Subjects were selected from a
sample of 33 s tudents at the University of Pittsburgh pre-
screened on the Harvard Group Scale of Hypnotic Susceptibil-
ity: Form A (Shor and Orne, 1962). Following a hypnotic
induction, high scorers (>8) were tested for their ability to
experience functional pain from an inactivated Medoc thermal
probe attached to the palm of the right hand. From that group,
five female and three male subjects, 21 to 50 years in age, who
reported consistent pain experience in the absence of stimulation
were selected for imaging with fMRI. Subjects were hypnotized
upon entering the MR scanner using an induction described in
detail elsewhere (Whalley and Oakley, 2003), and all experi-
mental procedures were carried out fol lowing the hypnotic
induction. After each scanning, block verbal ratings were taken
concerning the intensity of the six previous pain experiences.
Additional deepening instructions were provided to subjects
immediately following the feedback of ratings, and reinforce-
ment of the pain suggestion was provided before the initiation
of each scanning block.
Imaging procedure
Brain activation was inferred based on measurement of the
blood oxygen level-dependent (BOLD) contrast (Ogawa et al.,
1990). These measurements were acquired at 3 T using a reverse
spiral technique (TE = 25 ms, TR = 1.5 s, flip angle = 60j,64

64
matrix) described in detail elsewhere (Noll et al., 1995; Stenger et
al., 2000). As in the evaluation procedure, the subjects were fitted
with the thermal stimulator and told to expect noxious heat pulses to
the palmar surface of their right hand interspersed by 30-s rest over
6 min. A single tap to the foot indicated arrival of the stimulus, and
two taps indicated the beginning of the rest. Crucially, actual
Fig. 1. Graph of the average pain ratings for the physically induced and
hypnotically induced pain experience with standard deviations shown as
error bars. Physically induced pain resulted in significantly greater pain
ratings (P < 0.001).
S.W.G. Derbyshire et al. / NeuroImage 23 (2004) 392–401 393

noxious heat pulses (48.5jC) were delivered following only three of
the six single taps. The other three single taps and all six double taps
were accompanied by non-noxious heat (37.0jC). Functional data
were collected in two blocks of 6 min each to derive 3 min of
physically induced pain, 3 min of hypnotically induced pain, and 6
min of rest. The conditions (physically and hypnotically induced
pain) were alternated within blocks, and the alternation was
counterbalanced across blocks and subjects.
Fig. 2. Activated voxels during physically induced pain (left, red-yellow scale), hypnotically induced pain (middle, blue-purple scale), and the imagined
condition (right, yellow-green scale). The effects are shown as SPMs superimposed on an averaged structural MRI derived from the subject’s own structural
scans. At the top are sagittal slices 6 and 2 mm lateral to the midline. Below are coronal slices 20 mm posterior (negative), on (0 mm), and 12 mm anterior
(positive) to the anterior commissure. At the bottom are the surface projections. Regions of interest are numbered and significance detailed in tables 1 – 3.
S.W.G. Derbyshire et al. / NeuroImage 23 (2004) 392–401394

A further block of functional data was collected with
instructions to imagine the heat increasing to a painful level
following a single tap to the foot. Subjects were explicitly told
that the thermal probe would not be activated during this block
and that they were to simply imagine the heat pain as clearly as
possible following a single tap and to imagine the probe
becoming deactivated following two taps. The probe tempera-
ture remained at 37.0jC throughout. This block derived 3 min
of imagined pain and 3 min of rest.
Data analysis
Data analysis was performed using SPM2 (Wellcome Trust
Centre for the Study of Cognitive Neurology), described in
detail elsewhere (Friston et al., 1995). In summary, head
movement between scans was corrected by aligning all subse-
quent scans with the first. Each realigned set of scans from
every subject was coregistered with his or her own hi-res
structural MRI image and reoriented into the standardized
anatomical space of the average brain provided by the Montreal
Neurological Institute (MNI). To increase the signal to noise
ratio and accommodate variability in functional anatomy, each
image was smoothed in X, Y,andZ dimensions with a
Gaussian filter of 10 mm (FWHM). For each subject, a box-
car model with a hemodynamic delay function was fitted to
each voxel to contrast the effects of interest with rest gener-
ating a statistical parametric map that was then assessed for
significance at the second level for the group analysis shown in
Fig. 2. Baseline drifts were removed by applying a high-pass
filter and any artifact from the motion correction removed by
applying the correction parameters as covariates of no interest.
The random effects implementation c orrects for variability
between subjects so that outlying subjects cannot drive the
result. Brain regions with a large stat istic correspond to
structures whose BOLD response shares a substantial amount
of variance with the conditions of interest. Images were thresh-
olded at an arbitrary P < 0.01 with an extent threshold of 50
contiguous voxels. Directed searches of activation were con-
ducted on the thalamus, insula, S1, S2, and mid- and peri-
genual anterior cingulate ( pACC), prefrontal, and i nferior
parietal cortices. The multiple comparisons problem of simul-
taneously assessing all the voxel statistics was addressed via
correction for the total number of voxels reported active using
the false discovery rate (Genovese et al., 2001), via a correc-
tion for voxels within a region of interest or spherical volume
of 12-mm diameter centered upon the search region, or via the
cluster threshold (Friston et al., 1994). These methods are
consistent with those adopted elsewhere (Derbyshire, 2000;
Derbyshire et al., 1997, 2002; Faymonville et al., 2003; Rain-
ville et al., 1997) and provide a reasonable balance of protec-
tion against false-positive without artificially concealing the real
profile of activation.
Results
Subjects rated the perceived intensity of each physically
induced (PI) and hypnotically induced (HI) stimulus immediately
following each scanning block using a verbal rating scale (0, no
pain; 10, maximal pain). Average pain rating following actual
delivered stimulation (PI) was 5.7 (range 3–10), and average
rating without stimulation (HI) was 2.8 (range 1 –9) and is
illustrated in Fig. 1. This difference was statistically significant
( P < 0.001). All of the subjects confirmed that they imagined
the pain clearly in the imagined block, and only one reported
actually experiencing pain (of a low intensity and only on some
trials) in this condition. Four of the subjects reported a sensation
of increased heat in the imagining condition. As no pain was
actually expected in this condition, pain ratings were not
solicited.
The profile of brain activation dependent upon these percep-
tual changes in pain intensity is illustrated in Fig. 2 and
tabulated in Table 1. Activation of the thalamus, anterior
cingulate cortex (A24V/32V), cerebellum, S2, insula, inferior pari-
etal cortex [Brodmann area (BA) 39/40], and prefrontal cortex
(BA 9/10/46) are common to both physically and hypnotically
induced pain, although generally with greater i ntensity and
extent during actual s timulation. The imagined condition, in
contrast, provided minimal activation in the ACC (A32V extend-
ing into medial premotor cortex), insula, and S2. Activation in
S1 was observed only during HI pain.
The differences in activation betwe en these conditions were
formally assessed and the results shown in Fig. 3 and Tables 2 and
3. HI pain resulted in marginally greater activity of the midinsula,
S1, and orbitofront al cortex (BA 11/47), while actual noxious
stimulation produced greater activity of the thalamus and mid-
(A24V) and perigenual anterior cingulate (A24), and prefrontal and
inferior parietal cortices.
Greater activation throughout the pain matrix was evident for
both hypnotically and physically ind uced pai n relative to the
imagined condition.
To directly assess the dependence of brain activation upon
pain rating, the subjects with the highest and the lowest pain
ratings during HI pain were analyzed separately, and the result is
shown in Fig. 4. A subject with a matching average pain rating
during actual stimulation was also analyzed separately for com-
parison. As might be predicted from previous work (Coghill et
al., 2003; Derbyshire et al., 1997), higher subjective ratings are
associated with greater cerebral activity. Critically, this effect is
comparable whether the pain source is noxious heat or hypnotic
suggestion.
Discussion
fMRI data were obtained during conditions of physically
and hypnotically induced experiences of heat pain interleaved
with periods of rest, revealing common activation of the
thalamus, ACC, midanterior insula, and parietal and prefrontal
cortices (see Table 1 and Fig. 2). These findings demonstrate
the efficacy of suggestion following hypnotic induction in
producing altered sensory experience, as has been demonstrat-
ed elsewhere, with specificity of the response to the stimulus
under investigation (Faymonville et al., 2003; Rainville et al.,
1997). Compared to the rest condition, pain from a nocicep-
tive source and hypnotically induced pain both activated
regions of the brain that have been variously described as
belonging to a pain network or neuromatrix (Casey, 1999;
Derbyshire, 2000; Peyron et al., 2000; Price, 2000; Treede et
al., 1999).
In contrast, imagining the presence of a noxious heat stimulus
resulted in only minimal activation of the pain network, extensive-
S.W.G. Derbyshire et al. / NeuroImage 23 (2004) 392–401 395

Table 1
Regions with increased BOLD relative to rest due to HI, PI, and imagined conditions separately
BOLD increases relative to rest
(x, y, z coordinates) (region) (x, y, z coordinates) (x, y, z coordinates) (region)
HI T score P
FDRcorr
Cluster size Pcorr PI T score P
FDRcorr
Cluster Size Pcorr Imagined T score P
FDRcorr
Cluster size Pcorr
1. Thalamus
(0, 16, 0) 5.7 0.10
b
74 ns (18, 14, 10) 7.0 0.04
a
9590 0.00 No response – – – –
(8, 0, 4) 3.9 0.05
a
9590 0.00 No response – – – –
2. ACC
(4, 4, 48) 4.2 ns 417 0.05 (6, 10, 46)) 6.3 0.06 9590 0.00 No response – – – –
(6, 8, 34) 4.3 ns 417 0.05 (8, 20, 32) 12.0 0.04 9590 0.00 No response – – – –
3. pACC
No response – – – – No response – – – – No response – – – –
No response – – – – No response – – – – No response – – – –
4. Cerebellum
(12, 40, 30) 5.3 0.05
b
187 ns No response – – – – No response – – – –
(8. 46, 10) 5.6 0.05
b
118 ns (14, 72, 16) 9.8 0.04 872 0.00 No response – – – –
5. S1
(30, 16, 60) 9.7 0.01
b
332 ns No response – – – – No response – – – –
No response – – – – No response – – – – No response – – – –
6. S2/insula
(56, 16, 2) 9.7 0.01
a
808 0.00 (58, 28, 12) 6.7 0.06 110 ns (56, 8, 14) 5.6 0.04
b
339 ns
No response – – – – No response – – – – No response – – – –
S.W.G. Derbyshire et al. / NeuroImage 23 (2004) 392–401396

7. M. insula/putamen
(30, 0, 6) 5.7 0.05
b
374 ns (38, 2, 18) 16.1 0.03 9590 0.00 No response – – – –
(28, 10, 4) 10.3 0.00
a
2572 0.00 No response – – – – (34, 12, 8) 4.5 0.05
b
248 ns
8. A. insula
No response – – – – (34,12, 8) 9.1 0.04 9590 0.00 No response – – – –
(40, 18, 16) 7.4 0.03
a
2572 0.00 (38, 14, 6) 5.6 0.05 9590 0.00 No response – – – –
9. Inf. parietal cortex
(46, 48, 60) 7.5 0.04
a
1017 0.00 (32, 52, 40) 6.7 0.06 1120 0.00 No response – – – –
(56, 42, 48) 7.0 0.07
a
606 0.01 (30, 70, 56) 5.8 0.07 1452 0.00 No response – – – –
10. PFC (BA 9/46)
(50, 42, 24) 3.9 ns 56 ns (44, 34, 34) 15.1 0.03 9590 0.00 No response – – – –
(60, 14, 20) 8.3 0.01
b
2572 0.00 No response – – – – No response – – – –
11. PFC (BA 10/46)
No response – – – – (48, 42, 18) 10.0 0.04 9590 0.00 No response – – – –
(40, 60, 4) 7.7 0.01
b
2572 0.00 (40, 54, 2) 7.2 0.05 9590 0.00 No response – – – –
The areas are tabulated in terms of the brain region, as illustrated in Fig. 2, and their Brodmann areas (BA). The x, y, z coordinates plot each peak (defined as the pixel with the highest T score within each labeled
region) according to the MNI coordinate system (negative is left, posterior, and inferior; contralateral listed first for each region). P values are based on the false discovery rate (FDR)—see text for details. If a
region reached significance for any comparison, then the region is tabulated for all comparisons and for both sides except where no voxels reached the display threshold (P < 0.01 uncorrected) indicated as no
response. ACC indicates anterior cingulate cortex; pACC, perigenual anterior cingulate cortex; S1, primary sensory cortex; S2, secondary somatosensory cortex; M., mid; A., anterior; P., posterior; Inf., inferior;
PFC, prefrontal cortex.
P values are corrected for the whole brain based on the false discovery rate (FDR) except where indicated.
a
indicates correction made for the region wide voxels.
b
indicates correction applied for 925 voxels within a spherical volume of 12 mm diameter. See text for details.
S.W.G. Derbyshire et al. / NeuroImage 23 (2004) 392–401 397

ly reduced compared wit h both physically and hypnotically
induced pain experience. These results are com- parable to those
demonstrated using auditory sensation where physically presented
and hypnotically hallucinated sounds resulted in activation of the
right ACC, but imagining the same sound in hypnosis did not
(Szechtman et al., 1998).
Although we used hypnosis here as a tool to produce the
intended subjective effect, it is possib le to interpret the pain
experienced during the HI condition in terms of phenomena other
than hypnosis per se, such as a form of conditioned response to the
tap. In our experience, however, only highly hypnotizable individ-
uals are able to routinely and repeatedly report hallucinated expe-
rience, such as the presence of pain in the absence of a stimulus, and
high hypnotizables were directly selected for the current study.
Nevertheless, we cannot be certain of the extent to which the
hypnosis was responsible for creating the experience of pain
until further studies with and without hypnotic induction are
completed. Other research indicates that the hypnotic induction
may be neither necessary nor sufficient to produce response to
suggestion (Braffman and Kirsch, 1999), and nonhypnotic
suggestion has been used to produce functional headaches in
normal subjects (Schweiger and Parducci, 1981). None of this
materially alters the interpretation of our findings. Activation
observed during the hypnotically induced pain experience can
be interpreted without the usual caveats concerning incidental
sensory or motor processing that might be associated with an
actual stimulus regardless of the precise influence of hypnosis
in our study.
Although similar patterns were seen in the two conditions,
higher levels of activation were found with physically induced
pain compared with hypnotically induced pain in contralateral
thalamus, ACC, and orbitofrontal cortex and in the ipsilateral
parietal cortex. These larger responses can most easily be
explained as being due to the more intense pain experience
during PI but may also reflect the presence of peripheral
sensory information (Coghill et al., 2003; Derbyshire et al.,
1997).
Greater activation in the HI relative to PI condition incorpo-
rated bilateral S1 [overlapping with adjacent primary motor
cortex (M1)] partly as a consequence of decreased response in
the PI condition (decreases not shown). Variable S1 responses to
noxious stimuli have been reported with a mix of both increases
and decreases (Derbyshire, 2000; Derbyshire et al., 1997; Peyron
et al., 2000). In general, S1 activation occurs in about 50% of
pain studies and is usually within the appropriate somatotopical
region (Derbyshire et al., 1997). Regions of S1 not currently
engaged by the stimulus (such as the foot area when stimulating
the hand) have been demonstrated as reducing blood flow
possibly to enhanc e the spatial localization of the stimulus
(Apkarian et al., 1992; Drevets et al., 1995). These spatial
localization mechanisms may be more apparent when delivering
an actual stimulus relative to the hypnotically induced pain
experience.
Significant activation in the PI condition relative to HI also
incorporates the perigenual ACC (pACC, A24 approachin g
A25). This effect follows de creased resp onse in the HI
condition. Decreased pACC activation has been previously
reported during the anticipa tory phase bef ore delivery o f
stimulation that may be similar to the anticipation or internal
monitoring of sensory information during HI (Porro et al.,
2002).
Overall, however, Fig. 2 illustrates a considerable similarity
in the processing of both hypnotically and physically induced
pain but not with imagined pain. Fig. 4 further demonstrates
predictable activation based on the perceptual report of pain
experience independent of ac tual no cicep tive input . Thes e
findings extend beyond the general suggestion of a neural
network for pain by providing direct evidence that regional
activation is specifically and actively involved in the generation
of pain in the absence of stimulation. To our knowledge,
this is the first demonstration of a functional pain experience
Fig. 3. Differences between the physically induced (PI) and hypnotically
induced (HI) conditions to the left and the differences compared with the
imagined (IMA) condition to the right. Image layout is as for Fig. 1.
S.W.G. Derbyshire et al. / NeuroImage 23 (2004) 392–401398

measured with brain imaging in healthy normal controls. Direct
evidence for such a mechanism be ing present in c linical
functional pain must await further studies with chronic pain
patients. Nevertheless, by providing a material basis for
pain experience in the absence of injury or other physical
stimulus, these findings support the possibility of direct cortical
Table 2
Regions with significantly greater BOLD response during HI compared with PI (HI > PI) or vice versa (PI > HI)
Differences between HI and PI
HI > PI T score P
FDRcorr
Cluster size Pcorr PI > HI T score P
FDRcorr
Cluster size Pcorr
1. Thalamus, no difference – – – – (14, 14, 10) 11.05 0.02
a
1474 0.00
2. mACC, no difference – – – – (4, 26, 34) 9.4 0.03
a
1474 0.00
3. pACC, no difference – – – – (8, 36, 12) 8.4 0.03
b
1474 0.00
4. Cerebellum, no difference – – – – No difference – – – –
5. S2, no difference – – – – No difference – – – –
6. M. insula (38, 2, 16) 6.6 0.06
b
84 ns No difference – – – –
7. A. insula, no difference – – – – No difference – – – –
8. S1 (50, 14, 42) 5.6 ns 410 0.03 No difference – – – –
9. OFC (42, 34, 12) 6.4 0.05
b
63 ns PFC (42, 30, 34) 5.1 0.05
b
160 ns
10. IPC, no difference – – – – (32, 72, 54) 12.5 0.00
b
1149 0.00
The areas are tabulated in terms of the brain regions as illustrated in Fig. 3. OFC indicates orbitofrontal cortex; IPC, inferior parietal cortex. Other details and
abbreviations are as for Table 1.
Table 3
Regions with significantly greater BOLD response during HI compared with the imagined condition (HI > Imagined) and during PI compared with the
imagined condition (PI > imagined)
Comparisons with imagined
HI > Imagined T score P
FDRcorr
Cluster size Pcorr PI > Imagined T score P
FDRcorr
Cluster size Pcorr
1. Thalamus
(16, 24, 2) 5.6 0.03
b
2050 0.00 (12, 10, 14) 6.1 0.05
a
839 0.00
(14, 14, 0) 7.2 0.03
b
2050 0.00
2. mACC
(4, 10, 46) 5.8 0.04
b
471 0.02 (8, 24, 32) 6.9 0.05
a
789 0.00
(10, 30, 26) 7.1 0.04
a
789 0.00
3. pACC
No difference – – – – No difference – – – –
4. Cerebellum
(6, 46, 12) 10.3 0.00
b
2050 0.00 (26, 60, 24) 7.6 0.01
b
1023 0.00
(18, 70, 22) 10.5 0.00
b
1023 0.00
5. S2 S2/ P. insula
(68, 4, 4) 8.7 0.01
b
2708 0.00 (42, 12, 18) 6.3 0.04
b
921 0.00
6. P. insula M. insula
(36, 26, 8) 6.1 0.03
b
2708 0.00 (34, 2, 16) 10.2 0.01
b
921 0.00
7. A. insula
No difference – – – – (36, 16, 12) 4.3 0.05
b
921 0.00
(38, 14, 6) 5.3 0.04
b
2083 0.00
8. S1
(34, 28, 60) 10.2 0.02
a
2708 0.00 No difference – – – –
9. PFC
(48, 28, 14) 4.2 0.08
b
156 ns (46, 34, 34) 8.6 0.01
b
495 0.04
(42, 58, 4) 7.5 0.01
b
192 ns (32, 60, 12) 6.9 0.04
b
2083 0.00
10. IPC
(38, 56, 42) 4.7 0.03
b
330 ns (28, 50, 44) 7.7 0.06
a
774 0.00
(30, 50, 44) 9.0 0.03
a
1845 0.00
The areas are tabulated in terms of the brain regions as illustrated in Fig. 3. P. Insula indicates posterior insula. Other details and abbreviations are as for
Table 1.
S.W.G. Derbyshire et al. / NeuroImage 23 (2004) 392–401 399

involvement in the generation of some clinical functional pain
disorders.
Acknowledgments
This work was supported in part by a seed grant from the
University of Pittsburgh Department of Anesthesiology (UPP05-
862501-50540-999890) and by pilot funding from the University
of Pittsburgh Department of Radiology. MGW is supported by a
studentship from the Department for Work and Pensions (UK). His
participation in this project was assisted by a generous contribution
from the Bogue Fellowship. We thank Denise Davis for assistance
and technical advice and Simon Wessely for comments on an
earlier draft of this manuscript.
References
Apkarian, V.A., Stea, R.A., Manglos, S.H., Szevernyi, N.M., King, R.B.,
Thomas, F.D., 1992. Persistent pain inhibits contralateral somatosenso-
ry cortical activity in humans. Neurosci. Lett. 140, 141 – 147.
Barber, T.X., Hahn, K.W., 1964. Experimental studies in ‘‘hypnotic’’ be-
haviour: physiologic and subjective effects of imagined pain. J. Nerv.
Ment. Dis. 139, 416 – 425.
Bennett, R.M., Cook, D.M., Clark, S.R., Burckhardt, C.S., Campbell, S.M.,
1997. Hypothalamic-pituitary-like growth factor-I axis dysfunction in
patients with fibromyalgia. J. Rheumatol. 24, 1384 – 1389.
Braffman, W., Kirsch, I., 1999. Imaginative suggestibility and hypnotiz-
ability: an empirical analysis. J. Pers. Soc. Psychol. 77, 578 – 587.
Casey, K.L., 1999. Forebrain mechanisms of nociception and pain:
analysis through imaging. Proc. Natl. Acad. Sci. U. S. A. 96,
7668 – 7674.
Coghill, R.C., McHaffie, J.G., Yen, Y.-F., 2003. Neural correlates of inter-
individual differences in the subjective experience of pain. Proc. Natl.
Acad. Sci. U. S. A. 100, 8538 – 8542.
Croft, P., 2000. Testing for tenderness: what’s the point? J. Rheumatol. 27,
2531 – 2533.
Derbyshire, S.W.G., 1999. The IASP definition captures the essence of pain
experience. Pain Forum. 8, 106 – 109.
Derbyshire, S.W.G., 2000. Exploring the pain ‘‘neuromatrix’’. Curr. Rev.
Pain 6, 467 – 477.
Derbyshire, S.W.G., 2004. The biopsychosocial model: meaningless catch-
phrase or fundamental cornerstone? J. Cancer Pain (In press).
Derbyshire, S.W.G., Jones, A.K.P., Devani, P., Friston, K.J., Feinmann, C.,
Harris, M., Pearce, S., Watson, J.D.G., Frackowiak, R.S.J., 1994. Ce-
rebral responses to pain in patients with atypical facial pain measured
by positron emission tomography. J. Neurol. Neurosurg. Psychiatry 57,
1166–1172.
Derbyshire, S.W.G., Jones, A.K.P., Gyulai, F., Clark, S., Townsend, D.,
Firestone, L., 1997. Pain process ing dur ing three levels of noxious
stimulation produces differential patterns of central activity. Pain 73,
431 – 445.
Derbyshire, S.W.G., Jones, A.K.P. , Creed, F., Starz, T., Meltzer, C.C.,
Townsend, D.W., Peterson, A.M., Firestone, L., 2002. Cerebral
responses to noxious thermal stimulation in chronic low back pain
patients and normal controls. NeuroImage 16, 158 – 168.
Drevets, W.C., Burton, H., Videen, T.O., Snyd er, A.Z., Simpson, J.R.,
Raic hle, M. E., 1995. Blood flow changes in human somatosensory
cortex during anticipated stimulation. Nature 373, 249 – 252.
Dudley, D.L., Holmes, T.H., Martin, C.J., Ripley, H.S., 1966. Hypnotically
induced facsimile of pain. Arch. Gen. Psychiatry 15, 258 – 265.
Faymonville, M.E., Roediger, L., Del Fiore, G., Delgueldre, C., Phillips,
C., Lamy, M., Luxen, A., Maquet, P., Laureys, S., 2003. Increased
cerebral functional connectivity underlying the antinociceptive effects
of hypnosis. Cognit. Brain Res. 17, 255 – 262.
Friston, K.J., Worsley, K.J., Frackowiak, R.S.J., Mazziotta, J.C., Evans,
A.C., 1994. Assessing the significance of focal activations using their
spatial extent. Hum. Brain Mapp. 1, 214 – 220.
Fig. 4. Results from three individual subjects. At the bottom are activations during HI pain for the subject with the lowest average pain rating (average rating of
six HI = 1), in the middle are results for the subject with the highest average rating (=5) for HI, and at the top are results from a single subject with a PI rating
matching the highest HI rating (=5). The SPM results are shown superimposed on a left, contralateral, sagittal slice, a coronal slice, and projected onto the left
surface of each subject’s own brain.
S.W.G. Derbyshire et al. / NeuroImage 23 (2004) 392–401400

Friston, K.J., Holmes, A.P., Worsley, K.J., Poline, J.P., Frith, C.D., Frack-
owiak, R.S.J., 1995. Statistical parametric maps in functional imaging: a
general approach. Hum. Brain Mapp. 2, 189 – 210.
Genovese, C.R., Lazar, N.A., Nichols, T.E., 2001. Thresholding of statis-
tical maps in functional neuroimaging using the false discovery rate.
NeuroImage 15, 870 – 878.
Gracely, R.H., Petzke, F., Wolf, J.M., Clauw, D.J., 2002. Functional mag-
netic resonance imaging evidenc e of augmented pain processing in
fibromyalgia. Arthritis Rheum. 46, 1333 – 1343.
Halligan, P.W., Athwal, B., Oakley, D.A., Frackowiak, R.S.J., 2000. Im-
aging hypnotic paralysis: implications for conversion hysteria. Lancet.
355, 986 – 987.
Hilgard, E.R., Morgan, A.H., Lange, A.F., Lenox, J.R., Macdonald, H.,
Marshall, G.D., Sa chs, L.B., 1974. Heart r ate changes in pain and
hypnosis. Psychophysiology 11, 692 – 702.
Marshall, J.C., Halligan, P.W., Fink, G.R., Wade, D.T., Frackowiak, R.S.J.,
1997. The functional anatomy of a hysterical paralysis. Cognition. 64,
B1 – B8.
Mountz, J.M., Bradley, L.A., Modell, J.G., Alexander, R.W., Triana-Alex-
ander, M., Aaron, L.A., Stewart, K.E., Alarcon, G.S., Mountz, J.D.,
1995. Fibromyalgia in women: abnormalities of regional cerebral blood
flow in the thalamus and the caudate nucleus are associated with low
pain threshold levels. Arthritis Rheum. 38, 926 – 938.
Noll, D.C., Cohen, J.D., Meyer, C.H., Schneider, W., 1995. Spiral K-space
MR imaging of cortical activation. J. Magn. Reson. Imaging 5, 49 – 56.
Oakley, D.A., 1999. Hypnosis and conversion hysteria: a unifying model.
Cogn. Neuropsychiatry 4, 243 – 265.
Oakley, D.A., Ward, N.S., Halligan, P.W., Frackowiak, R.S.J., 2003. Dif-
ferential brain activations for malingered and subjectively ‘real’ paral-
ysis. In: Halligan, P.W., Bass, C., Oakley, D.A. (Eds.), Malingering and
Illness Deception. Oxford Univ. Press, Oxford, pp. 267 – 284.
Ogawa, S., Lee, T.M., Kay, A.R., Tank, D.W., 1990. Brain magnetic res-
onance imaging with contrast dependent on blood oxygenation. Proc.
Natl. Acad. Sci. U. S. A. 87, 9868 – 9872.
Peyron, R., Laurent, B., Garcia-Larrea, L., 2000. Functional imaging of
brain responses to pain. A review and meta-analysis. Neurophysiol.
Clin. 30, 263 – 288.
Porro, C.A., Baraldi, P., Pagnoni, G., Serafini, M., Facchin, P., Maieron,
M., Nichelli, P., 2002. Does anticipation of pain affect cortical noci-
ceptive systems? J. Neurosci. 22, 3206 – 3214.
Price, D.D., 2000. Psychological and neural mechanisms of the affective
dimension of pain. Science 288, 1769 – 1772.
Rainville, P., Duncan, G.H., Price, D.D., 1997. Pain affect encoded in
human anterior cingulate but not somatosensory cortex. Science 277,
968 – 971.
Raz, A., Shapiro, T., 2002. Hypnosis and neuroscience: a cross talk between
clinical and cognitive research. Arch. Gen. Psychiatry 59, 85 – 90.
Russell, I.J., Orr, M.D., Littman, B., Vipraio, G.A., Alboukrek, D., Micha-
lek, J.E., Lopez, Y., MacKillip, F., 1994. Elevated cerebrospinal fluid
levels of substance P in patients with fibromyalgia syndrome. Arthritis
Rheum. 37, 1593 – 1601.
Schweiger, A., Parducci, A., 1981. Nocebo: the psychologic induction of
pain. Pavlovian J. Biol. Sci. 16, 140 – 143.
Shor, R.E., Orne, E.C., 1962. Harvard Group Scale of Hypnotic Suscepti-
bility, Form A Consulting Psychologists Press, Palo Alto, CA.
Stenger, V.A., Boada, F.E., Noll, D.C., 2000. Three-dimensional tailored
RF pulses for the reduction of susceptibility artifacts in T2*-weighted
functional MRI. Magn. Reson. Med. 44, 525 – 531.
Szechtman, H., Woody, E., Bowers, K.S., Nahmias, C., 1998. Where the
imaginal appears real: a positron emission tomography study of audi-
tory hallucination. Proc. Natl. Acad. Sci. U. S. A. 95, 1956 – 1960.
Treede, R.D., Kenshalo, D.R., Gracely, R.H., Jones, A.K.P., 1999. The
cortical representation of pain. Pain 79, 105 – 111.
Ward, N.S., Oakley, D.A., Frackowiak, R.S.J., Halligan, P.W., 2003. Dif-
ferential brain activations during intentionally simulated and subjective-
ly experienced paralysis. Cogn. Neuropsychiatry 8, 295 – 312.
Wessely, S., Nimnuan, C., Sharpe, M., 1999. Functional somatic syn-
dromes: one or many? Lancet. 354, 936 –939.
Whalley, M.G., Oakley, D.A., 2003. Psychogenic pain: a study using mul-
tidimensional scaling. Contemp. Hypn. 20, 16 – 24.
S.W.G. Derbyshire et al. / NeuroImage 23 (2004) 392–401 401