1257

Anesthesiology
2000; 92:1257– 67
© 2000 American Society of Anesthesiologists, Inc.
Lippincott Williams & Wilkins, Inc.

Neural Mechanisms of Antinociceptive Effects of

Hypnosis
Marie Elisabeth Faymonville, M.D., Ph.D.,* Steven Laureys, M.D., Ph.D.,† Christian Degueldre, Ph.D.,‡
Guy Del Fiore, Ph.D.,§ André Luxen, Ph.D.,储 Georges Franck, M.D., Ph.D.,# Maurice Lamy, M.D., Ph.D.,**
Pierre Maquet, M.D., Ph.D.††

Background: The neural mechanisms underlying the modu- that would be more or less activated in hypnosis than in control
lation of pain perception by hypnosis remain obscure. In this conditions, under noxious stimulation).
study, we used positron emission tomography in 11 healthy Results: Hypnosis decreased both pain sensation and the
volunteers to identify the brain areas in which hypnosis mod- unpleasantness of noxious stimuli. Noxious stimulation caused
ulates cerebral responses to a noxious stimulus. an increase in regional cerebral blood flow in the thalamic
Methods: The protocol used a factorial design with two fac- nuclei and anterior cingulate and insular cortices. The hypnotic
tors: state (hypnotic state, resting state, mental imagery) and state induced a significant activation of a right-sided extrastriate
stimulation (warm non-noxious vs. hot noxious stimuli applied area and the anterior cingulate cortex. The interaction analysis
to right thenar eminence). Two cerebral blood flow scans were showed that the activity in the anterior (mid-)cingulate cortex
obtained with the 15O-water technique during each condition. was related to pain perception and unpleasantness differently
After each scan, the subject was asked to rate pain sensation and in the hypnotic state than in control situations.
unpleasantness. Statistical parametric mapping was used to de- Conclusions: Both intensity and unpleasantness of the nox-
termine the main effects of noxious stimulation and hypnotic ious stimuli are reduced during the hypnotic state. In addition,
state as well as state-by-stimulation interactions (i.e., brain areas hypnotic modulation of pain is mediated by the anterior cingu-
late cortex. (Key words: Functional neuroimaging; pain; statis-
tical parametric mapping.)
This article is featured in “This Month in Anesthesiology.”
䉫 Please see this issue of ANESTHESIOLOGY, page 5A.

HYPNOSIS combined with slight conscious intravenous

* Associate Professor, Department of Anesthesiology and Intensive
sedation (hypnosedation) and local anesthesia offers a
Care Medicine. valuable alternative to traditional general anesthesia.1– 4
† Research Fellow, Cyclotron Research Centre and Department of In our center, the technique has been used in more than
Neurology. 1,800 surgical interventions since 1992. The effective-
‡ Research Fellow, Cyclotron Research Centre. ness of hypnosis in producing analgesia has been dem-
§ Senior Research Associate, Cyclotron Research Centre. onstrated by two clinical studies. A retrospective study
储 Professor, Cyclotron Research Centre. first showed that hypnosis as an adjunct procedure to
# Professor, Cyclotron Research Centre and Department of Neu- conscious intravenous sedation provides significant peri-
rology. operative pain and anxiety relief. These benefits were
** Professor, Department of Anesthesiology and Intensive Care obtained despite a significant reduction in drug require-
Medicine. ments.1 A prospective randomized study confirmed
†† Senior Research Associate, Fonds National de la Recherche Sci- these observations.2
entifique dé Belique, Belgium.
In a recent positron emission tomography (PET) study
Received from the Departments of Anesthesiology and Intensive aimed at differentiating cortical areas involved in pain
Care Medicine and Neurology, and the Cyclotron Research Centre,
University Hospital of Liège, Liège, Belgium. Submitted for publication
affect, Rainville et al.5 used hypnotic suggestions to alter
March 12, 1999. Accepted for publication November 4, 1999. Sup- selectively the unpleasantness of remained noxious stim-
ported by grant No. 3.4536.99 from the Fonds National de la Recher- uli, without changing the perceived intensity. In these
che Scientifique de Belgique, Belgium; by the Reine Elisabeth Medical conditions, anterior cingulate cortex (ACC) activity was
Foundation, Belgium; and by research funds of the University of Liège,
Liège, Belgium.
shown to be selectively correlated with unpleasantness.
Address reprint requests to Dr. Faymonville: Department of Anes-
However, our experimental design differed in that vol-
thesia and Intensive Care Medicine, University Hospital of Liège B35, unteers were asked to rate unpleasantness and perceived
4000 Liège, Belgium. Address electronic mail to: anesrea@ulg.ac.be intensity of noxious stimuli without a specific demand to

Anesthesiology, V 92, No 5, May 2000

1258

FAYMONVILLE ET AL.

maintain either one or the other constant. By this it is silent. The HS was considered to be present when roving
meant that subjects were not asked to actively induce eye movements were observed on oculography and if,
analgesia but only to recall pleasant life experiences, just before the scan, the subjects responded by a prear-
without any reference to pain perception.1,2 The ratio- ranged foot movement that he/she felt in the HS. Slow
nale of the present study was to explore the brain mech- ocular movements are regularly observed in the HS in
anisms underlying the modulation of pain perception isolation or intermingled with few saccades. This pattern
proper to our clinical hypnotic protocol. of ocular movements, in conjunction with the subject’s
behavior, was used to differentiate the HS from other
states. Polygraphic recordings ensured that no sleep oc-
Materials and Methods curred during the experimental session.
Each subject was scanned twice in both levels of stim-
Subjects ulation (non-noxious and noxious) in each of the three
This study was approved by the Ethical Committee of states (12 scans per subject). After each measurement,
the Faculty of Medicine of the University of Liège. the subjects were asked to verbally rate the noxious
Healthy right-handed drug-free unpaid volunteers were stimulus intensity and unpleasantness on a scale from 0
considered for selection after written informed consent
to 10 (for sensation, 0 ⫽ no pain sensation, 10 ⫽ most
was obtained. From a cohort of 30 screened subjects, 11
intense painful sensation imaginable; for unpleasantness,
(4 women, 7 men; mean age, 31.7 yr; age range, 27–55
0 ⫽ not at all unpleasant, 10 ⫽ most unpleasant imagin-
yr) were selected because they were scored as highly
able). To avoid multiple hypnotic inductions, the fifth to
hypnotizable subjects (score ⬎ 8 of 12) according to a
eighth scans were always made in HS. The order of the
French version of the Stanford Hypnotic Susceptibility
other two states, and of the non-noxious and noxious
Scale–Form C.6 During the selection procedure, which
stimulations, was pseudorandomized over subjects. Sub-
took place several weeks before the experimental ses-
jects were warned that scans started but were not told in
sion, detailed information about pleasant life experi-
ences that the subject wanted to use during the experi- which order the different stimulations would occur. Sub-
ment was obtained through a semistructured interview. jects were instructed to keep their eyes closed through-
out the experimental session. Ambient noise was re-
duced to a minimum, and ambient light was dimmed.
Experimental Design
Thermal Stimulation. Thermal stimuli were deliv-
Experimental Conditions. The experiment followed
ered by a Marstock thermal stimulator (Somedic: ther-
a factorial design with two factors: stimulation (warm
non-noxious vs. hot noxious) and state (resting state motest Type I; Senselab, Upsala, Sweden) that delivers
[RS], mental imagery [MI], hypnotic state [HS]). calibrated and reproducible thermal stimulations via a
In the first condition (RS), the subjects were asked to water-cooled probe (2.5 ⫻ 5 cm). The thermode was
empty their minds and remain immobile. In the second applied to the thenar eminence of the right hand. The
condition (MI), during the interscan interval, the sub- stimuli consisted of a ramp increase from 35°C to the
jects listened to sentences containing pleasant informa- predetermined level during 5 s, a plateau at this temper-
tion taken from their own past. Subjects were instructed ature for 5 s, and linear return to the baseline tempera-
to vividly imagine a pleasurable autobiographical mem- ture for 5 s. This sequence was repeated six times during
ory. The subjects were urged not to try to enter in the the scanning period. Thermal stimulation started 10 s
HS. During 90-s scanning periods, the experimenter re- before the second frame of the scans.
mained silent. Subjects confirmed by a foot movement Before the PET studies, target temperatures that were
that they used MI. In the third condition (HS), the sub- reproducibly experienced as warm and non-noxious
jects were scanned after the HS was induced. This con- (typically 39°C) or hot and noxious (typically 47°C)
dition started with a 3-min induction procedure involv- were carefully established for each subject before the
ing muscle relaxation. Subjects were then invited to study. Once established, these individual (non-noxious
reexperience their pleasant autobiographical memory. and noxious) temperatures were used during the corre-
As in clinical conditions, permissive and indirect sugges- sponding scans. Practice sessions were conducted so
tions were used to develop and deepen the HS. They that the anxiety and emotional reactions associated with
were continuously given cues for maintaining an HS. a novel experimental situation or unexpected noxious
However, during the scans, the experimenter remained stimuli would be reduced.

Anesthesiology, V 92, No 5, May 2000

 1259

PET STUDY OF PAIN MODULATION BY THE HYPNOTIC STATE

PET and Magnetic Resonance Imaging Acquisi- of interest: the pain ratings, the experimental states, and
tions. Before the scanning session, electrodes were put a covariate representing the interaction between ratings
in place to monitor electroencephalograph (C3–A2 and of pain perception and the states (the HS vs. control
C4 –A1), horizontal electrooculogram, and chin electro- states). The state regressor consisted of dummy variables
myogram. A venous catheter was inserted during local (⫺1 for RS and MI scans and 1 for the HS scans). The use
anesthesia in a left antebrachial vein. The subject’s head of pain ratings and states as regressors allowed the as-
was stabilized by a thermoplastic face mask secured to sessment of main effects of pain perception and the HS
the head holder (Truscan Imaging, Anapolis, MA). Ear- condition, respectively. These two covariates were cen-
phones were adapted to the subject’s head, and verbal tered, orthogonalized, and multiplied, element by ele-
communications were made at a distance via a micro- ment, to form the third covariate, which thus repre-
phone. Direct visual observation was maintained at all sented a state-by-stimulation interaction covariate. The
times. A transmission scan was performed to allow a
rationale of similar types of analysis was described by
measured attenuation correction. Twelve emission scans
Friston et al.11 In essence, this analysis looks for a differ-
were acquired at 8-min intervals in three-dimensional
ence in the slope of regression between cerebral blood
mode using a CTI 951 16/32 scanner (Siemens, Erlangen,
Germany). Each scan consisted of two frames: a 30-s flow (CBF) and pain ratings between the HS and the
background frame and a 90-s frame. The slow intrave- other states.
nous water (H215O) infusion was begun just before the In both types of analysis, the design matrix also in-
second frame to observe the head curve rising within the cluded the block effect as a confounding covariate.12
first 10 s of this frame. Six to eight millicuries (222–296 Global flow normalization was performed by propor-
MBq) were injected for each scan, in 10 ml saline, over tional scaling. Furthermore, the RS and MI were consid-
a period of 20 s. The infusion was totally automated so as ered together and contrasted to the HS. The collapse of
not to disturb the subject during the scanning periods. these states into a single one was considered when
Data were reconstructed using a Hanning filter (cutoff behavioral data showed no significant difference in pain
frequency: 0.5 cycle/pixel) and corrected for attenua- ratings between them (see Results).
tion and background activity. The resulting set of voxels for each contrast consti-
A high resolution (voxel size: 0.96 ⫻ 0.96 ⫻ 1.35 mm) tuted a map of the t statistic (SPM{t}). The SPM{t} were
T1-weighted structural magnetic resonance imaging scan then transformed to the unit normal distribution
was obtained for each subject on a 1.5 T imager (Mag- (SPM{z}). Whatever the analysis, the first step was to
netom, Siemens) a few days after the PET session. identify the main effects of pain and hypnosis. In these
contrasts, hypotheses existed as to which brain areas
PET Data Analysis should be found activated. Results were thus considered
Positron emission tomography data were analyzed us- significant at Z ⫽ 3.09 (P ⬍ 0.001, uncorrected). Based
ing the statistical parametric mapping software (SPM96 on previous literature, the main effect of noxious stim-
version; Wellcome Department of Cognitive Neurology, ulation was considered in upper midbrain, thalamic nu-
Institute of Neurology, London, United Kingdom7) im- clei, lentiform nuclei, primary and secondary somatosen-
plemented in MATLAB (Mathworks Inc., Sherborn, MA).
sory cortexes, the insula, and the ACC. On the basis of
In short, data from each subject were realigned using a
our previous study,13 the effect of hypnosis was sus-
least square approach and the first scan as a reference.8
pected to occur bilaterally in the occipital regions and
PET data were then coregistered to individual T1-
the ACC or on the left side in parietal, motor areas, and
weighted magnetic resonance imaging scans. After re-
alignement, all images were transformed into a standard the ventrolateral prefrontal cortex.
space8,9 and then smoothed using a 16-mm full width at However, the particular interest of the present study
half-maximum isotropic kernel. was in the state-by-stimulation interaction, looking for
Two separate statistical analyses were performed. The the brain areas that would be more (or less) activated by
first one was based on categoric comparisons, and the noxious stimulation during the HS than in other states.
second used a multiple regression approach. For cate- For this purpose, we considered the analysis as explor-
goric comparisons, the design matrix10 included the 12 atory and used a more conservative level of significance
conditions (scans) for each subject. For the regression (i.e., P ⬍ 0.05 corrected for multiple comparisons at the
analysis, the design matrix consisted of three covariates voxel level).10

Anesthesiology, V 92, No 5, May 2000

1260

FAYMONVILLE ET AL.

Results rating scale for unpleasantness did not differ from the
one for pain intensity. The interaction between state and
Behavioral Data thermal stimulation on ratings was significant [F(2,126) ⫽
The average temperature used for warm non-noxious 9.66; P ⬍ 0.001], demonstrating that subjects experienced
and noxious stimulation was, respectively, 39.1°C ⫾ 0.3 noxious stimulation differently when at rest, distracted,
and 47.2°C ⫾ 1.1 (mean ⫾ SD). or in the HS. A Tukey honest significant difference post
Figure 1 shows ratings of unpleasantness and pain hoc test showed that the state effect was only significant
sensation after thermal non-noxious and noxious stimu- for the HS versus RS (P ⬍ 0.001) and versus MI (P ⬍
lation in RS, MI, and HS. A three-way analysis of variance 0.001) but not for MI versus RS (P ⬎ 0.440).
with state (RS, MI, and HS) and thermal stimulation
(non-noxious vs. noxious) as independent factors, and PET Data
rating (unpleasantness vs. pain intensity) as within-sub- Categoric Comparisons. The SPM had 110 residual
ject variables, revealed no significant effect of the rating degrees of freedom, a smoothness estimate of 13.2 ⫻
variable [F(1,126) ⫽ 1.07; P ⬎ 0.30], indicating that the 14.3 ⫻ 14.7 mm and was composed of 193,799 voxels
(i.e., 553.6 resolution elements).
When all conditions were considered together, the
main effect of pain, as compared with non-noxious stim-
ulation, consisted of an activation in both thalamic nu-
clei (predominantly on the right side), in the right cau-
date nucleus, and in a region encompassing the left
insula and the ACC (fig. 2B and table 1). Other regions
that were not expected a priori were also significantly
activated: the right dorsolateral prefrontal cortex (Brodma-
nn’s area [BA] 8), and the orbitofrontal cortex on both
sides.
When the analysis concerned only “alert” states (RS
and MI), the main effect of noxious stimulation was
observed in the left insular cortex (fig. 2C and table 1).
The left orbitofrontal cortex was also activated, although
it was not included in our a priori hypotheses.
In the HS, activation was observed in response to
noxious stimulation in an area encompassing the ACC
(both BA 24 and 32), right caudate, left caudate, and left
putamen (fig. 2D and table 1). Further activation was
found in a region involving the right thalamus and ex-
tending caudally to the upper midbrain. Other regions
were also found activated but were not predicted a
priori: the right orbitofrontal cortex, the right dorsolat-
eral prefrontal cortex (BA 9), and the right inferior pari-
etal lobule (BA 40).
The comparison between the HS and the other two
states (RS and MI) showed activation in the right extra-
striate area (BA 19; fig. 3 and table 1). More anteriorly,
Fig. 1. Ratings of noxious sensation and unpleasantness during activated sites were present in the right ACC, one of
the three states (RS ⴝ resting state; MI ⴝ mental imagery; HS ⴝ which crossed the border between the ACC and the
hypnotic state). Note that hot noxious stimuli had higher rat-
ings than warm non-noxious ones. Ratings for noxious sensa- corpus callosum.
tion and unpleasantness are not significantly different from The state-by-stimulation interaction (table 1) looked
each other. For noxious hot stimuli, ratings are significantly for brain areas that would be more activated by hot
lower during the HS than during RS or MI, whereas RS and MI
ratings are not significantly different from each other. Boxes noxious (as compared with non-noxious) stimuli, in the
and whiskers represent, respectively, SEMs and SDs. context of the HS (as compared with RS and MI). This

Anesthesiology, V 92, No 5, May 2000

 1261

PET STUDY OF PAIN MODULATION BY THE HYPNOTIC STATE

Fig. 2. Categoric comparisons: main effect of noxious simulation. (A) The design matrix included 12 conditions (scans) for each
subject. (B) All conditions; (C) non-HS states; (D) HS. The results are displayed in a transparent brain normalized to the reference
space of Talairach and Tournoux,9 thresholded at P < 0.001.

analysis did not show any significant activation at the main effect of noxious stimulation was characterized by
chosen level for this contrast (P ⬍ 0.05, corrected for a significant activation of an area involving both thalami
multiple comparisons at the voxel level). However, at and caudate nuclei (fig. 4B and table 2). The left insula
the uncorrected level P ⬍ 0.001, a region across the ACC and the ACC were also found activated. Other (unex-
and corpus callosum (P ⫽ 0.13 at voxel level; Z ⫽ 4.25; pected) regions were found activated in the right orbito-
x ⫽ ⫺2 mm; y ⫽ 16 mm; z ⫽ 14 mm) as well as a medial frontal cortex, the right dorsolateral prefrontal cortex
polar prefrontal area (Z ⫽ 3.38; x ⫽ 0 mm; y ⫽ 60 mm; (BA 44/46 and 9), and left parietal cortex (BA 40). This
z ⫽ 26 mm) were found activated (not shown). No mode of analysis does not permit the separate evaluation
region was found less activated in the HS than in other of the effect of noxious stimulation in alert states and HS.
states during pain perception. Significant regression was found with the state covari-
Regression Analysis. The SPM had 118 residual de- ate in the ACC, indicating an increased CBF in these
grees of freedom, a smoothness estimate of 13.4 ⫻ 14.5 regions in the HS as compared with RS and MI (fig. 4C
⫻ 14.9 mm, and was composed of 193,799 voxels (i.e., and table 2). This activation area continued caudal to the
539.2 resolution elements). ventral striatum. The left caudate nucleus was also sig-
Using subjects’ pain sensation ratings as regressor, the nificantly activated.

Anesthesiology, V 92, No 5, May 2000

1262

FAYMONVILLE ET AL.

Table 1. Results from the Categorical Comparisons

Side Region x y z Z score

Increases in rCBF caused by noxious stimulation (all conditions)*

Left Insula ⫺28 14 10 5.16
Left Anterior cingulate cortex ⫺8 22 30 3.39
Right Thalamus 18 ⫺22 8 3.76
Left Thalamus ⫺10 ⫺26 8 3.34
Right Dorso-lateral prefrontal cortex 4 34 38 3.83
Right Orbito-frontal cortex ⫺24 38 ⫺24 3.93
Left Orbito-frontal cortex 24 32 ⫺24 4.91
Right Caudate nucleus 24 10 18 3.90
Increases in rCBF caused by noxious stimulation (R and MI)*
Left Insula ⫺30 12 8 4.61
Right Orbito-frontal cortex 24 32 ⫺24 3.99
Increases in rCBF caused by noxious stimulation (the HS alone)*
Anterior cingulate cortex (BA 24) ⫺2 18 22 4.52
Anterior cingulate cortex (BA 32) 2 28 22 3.77
Right Thalamus 12 ⫺14 0 3.77
Left Putamen ⫺24 16 8 3.67
Mesencephalon 4 ⫺28 ⫺8 3.64
Right Orbito-frontal cortex 34 36 ⫺22 3.74
Right Dorso-lateral prefrontal cortex 46 34 36 3.54
Right Inferior parietal lobule 50 ⫺58 42 3.27
Right Caudate nucleus 6 14 10 3.57
Left Caudate nucleus ⫺10 4 18 3.35
Increases in rCBF caused by the HS as compared to both R and MI state
Right Anterior cingulate cortex (BA 24) 8 34 2 3.73
Right Anterior cingulate cortex 18 14 24 3.52
Right Extrastriate cortex 50 ⫺74 ⫺10 3.51
Interaction state by stimulation†
Anterior cingulate cortex/corpus callosum ⫺2 16 14 4.25
Medial prefrontal cortex ⫺2 16 14 4.25

* In italics, the regions significant at P ⬍ 0.001 (uncorrected) that were not expected to be activated.
† In italics, the regions that were significant at P ⬍ 0.001 but did not survive correction for multiple comparisons at the voxel level (P ⬍ 0.05).

Finally, a significant interaction between pain sensa- DISCUSSION

tion ratings and state (fig. 4D and table 2) was observed
in a region involving the ACC (P ⫽ 0.047: Z ⫽ 4.51; BA Authenticity of HS
24; x ⫽ ⫺2; y ⫽ 18; z ⫽ 22). This region spreads rostral It is clear that our experimental protocol relies criti-
to area 32, reaching the vicinity of medial BA 9 and cally on the recognition of the HS and its differentiation
caudal toward the corpus callosum. The voxel with from control states, in particular MI. Three arguments
maximum Z value is located in the supracallosal part of corroborate the presence of the HS in our subjects dur-
the midcingulate cortex (fig. 5A). In the specific context ing scanning. First, the recording of slow ocular move-
of hypnosis, and in contrast to the control states, the ments has proven a valuable parameter in our clinical
ACC regional CBF increases proportionally to pain sen- and research protocols. These eye movements cannot be
sation (fig. 5B). Similar results were observed using pain willfully mimicked.14 At the very least, their recording
unpleasantness ratings. Again, no region was found less rules out the presence of a simulated state. Second, the
activated in the HS than in other states during the appli- subject’s behavior is characterized by an intense muscu-
cation of noxious stimuli. lar relaxation, a decrease in heart and respiratory rates,

Anesthesiology, V 92, No 5, May 2000

 1263

PET STUDY OF PAIN MODULATION BY THE HYPNOTIC STATE

lamic activation, although it may also be lacking.16 How-

ever, Adler et al.20 and Rainville et al.5 (quoted by Der-
byshire et al.17) found bilateral thalamic activation, the
latter with a reportedly ipsilateral predominance. Like-
wise, ipsilateral thalamic activation was observed for
mildly painful stimulation and not for more painful stim-
uli.16 The reason for this particular thalamic distribution
may also be related to the tonic mode of noxious heat
stimulation, as suggested by Derbyshire et al.17
When alert states are considered in isolation, the insu-
lar cortex controlateral to the noxious stimulation was
the only cortical area to be significantly activated. The
insular cortex is among the brain areas that are most
frequently reported as activated in response to noxious
Fig. 3. Categoric comparisons: main effect of state. Increases in
the hypnotic state (HS) as compared with the other states. The stimulation.16,17,19,21–23 More intriguing is the lack of
design matrix is the same as in Fig. 2. The results are displayed activation in other brain areas, in particular the thalamic
in a transparent brain normalized to the references space of
Talairach and Tournoux,9 thresholded at P < 0.001. nuclei and the ACC. This is in contrast to other reports
of functional neuroanatomy of the central processing of
and a sluggishness in verbal and motor response that are noxious stimuli.5,16,17 These negative results may be
more marked than at rest or during MI. In this respect, caused by various factors. Despite the restricted number
our subjects’ behavior corresponded to our clinical ob- of observations per subject in alert states (eight scans per
servation. Third, a statistically significant decrease in subject), a lack of statistical power is unlikely to be
pain ratings was observed during the HS only, a finding relevant here because there were 110 residual degrees of
that is in agreement with our clinical practice.2,15 Fur- freedom in our (categoric) design matrix. Furthermore,
thermore, our subjects testified that they were in the HS significant activation in ACC was found in the HS alone,
before each scan and confirmed their hypnotic experi- where the number of observations is even fewer (four
ence during debriefing. Each of these points, taken in scans per subject). We already pointed out the effect of
isolation, does not prove the presence of the HS in our a tonic, rather than phasic, noxious stimulation on the
subjects, but together they form a body of arguments regional CBF increases as detected by SPM. The intensity
that, by their cooccurrence, strongly suggest that this of the stimulation is also of importance. For instance, the
was indeed the case. thalamic nuclei and the ACC are not activated by “just
painful” stimuli but were activated by “moderately pain-
ful” stimulations.16 This factor is probably not relevant in
Main Effects of Noxious Stimulation
When all conditions are considered together, regional the present study because the target temperature for
CBF increases in response to noxious stimulation in non-noxious and noxious stimulations was set for each
various brain areas related to pain perception: thalamic subject before the scanning session. As indicated by
nuclei and anterior cingulate and insular cortices. These subjects’ ratings, the non-noxious and noxious stimuli
three sites are most commonly reported as activated could be easily discriminated. Finally, a carry-over of the
during noxious stimulation.16 We did not observe any antinociceptive effect of the HS during the post-HS con-
significant activation of somatosensory areas (SI and SII), trol scans remains possible. Indeed, pain ratings for
but these cortical regions are not systematically reported post-HS scans tended to be lower than pre-HS values,
in the literature.17 More specifically, the lack of activa- although this variation was not significant (e.g., for nox-
tion may be related to our mode of stimulation, which ious sensation, before HS: 5.9 ⫾ 2.2; after the HS: 5.3 ⫾
includes a tonic aspect and does not optimize SI/SII 2.3). In addition, in our clinical studies, postoperative
activation.17–19 pain was significantly lower in the hypnosis group de-
The thalamic activation, although bilateral, was pre- spite a standardized prescription of postoperative anal-
dominantly ipsilateral to the stimulation. Most studies of gesics.2 In these conditions, mixing pre-HS and post-HS
pain perception with PET reported contralateral tha- scans may have averaged out some regional activations.

Anesthesiology, V 92, No 5, May 2000

1264

FAYMONVILLE ET AL.

Fig. 4. Multiple regression analysis. (A) The design matrix included three covariates of interest: the pain ratings, the experimental
states, and a covariate representing the interaction between ratings of pain perception and the states (the hypnotic state vs. control
states). (B) Main effects of pain perception. (C) Main effect of state (increases in the hypnotic state as compared with the two other
states). (D) State-by-condition interaction. The results are displayed in a transparent brain normalized to the reference space of
Talairach and Tournoux9 thresholded at P < 0.001.

Main Effects of the HS subjects in the HS were verbally accompanied during the
We previously reported that the functional neuroanat- entire hypnotic session, including during the scanning
omy of the HS was characterized by the activation of a periods. The only instructions were to enter the HS and
widespread, mainly left-sided, set of cortical areas involv- let the HS imagery invade their consciousness. In the
ing occipital, parietal, precentral, premotor, ventrolat- present experiment, during the hypnotic session, the
eral prefrontal cortices, and a few right-sided regions: experimenter remained silent during the scanning peri-
occipital and anterior cingulate cortices.13 These results ods, and thermal stimuli were administered. It is proba-
were recently confirmed by another group.24 In the ble that, in these conditions, and although the subjects
present study, regional CBF distribution during the HS were not explicitly instructed to do so, most of the
differed from alert states only by a significant activation mentation in the HS was directed toward reducing pain
of a right-sided extrastriate area and the ACC. The differ- perception. This would explain the predominant activa-
ences in activation patterns are likely to be a result of the tion of the ACC, but we currently have no means to
experimental conditions. In our previous experiment, substantiate this.

Anesthesiology, V 92, No 5, May 2000

 1265

PET STUDY OF PAIN MODULATION BY THE HYPNOTIC STATE

Table 2. Results from the Regression Analysis

Side Region x y z Z score

Increases in rCBF due to pain ratings*

Left Insula ⫺30 10 16 4.94
Anterior cingulate (BA 32) ⫺2 26 30 4.08
Anterior cingulate (BA 24) ⫺6 12 30 3.23
Left Thalamus ⫺12 ⫺24 10 4.26
Right Thalamus 10 ⫺6 4 4.39
Right Orbito-frontal cortex 22 34 ⫺24 4.76
Right Dorso-lateral prefrontal cortex (BA 44/46) 62 18 22 3.62
Right Dorso-lateral prefrontal cortex (BA 9) 50 30 34 3.43
Left Parietal cortex (BA 40) ⫺56 ⫺54 44 3.64
Right Caudate nucleus 14 14 10 3.36
Left Caudate nucleus ⫺20 ⫺4 16 3.60
Increases in rCBF due to the HS as compared to both R and MI states*
Right Anterior cingulate cortex (BA 24) 8 34 6 3.89
Right Caudate nucleus 14 22 4 3.18
Left Caudate nucleus ⫺18 24 12 3.95
Interaction state by stimulation
Anterior cingulate cortex ⫺2 18 22 4.51

* In italics, the regions significant at P ⬍ 0.001 (uncorrected) that were not expected to be activated.

These results shed further light on brain function in clear. To explore the neural network that the ACC might
the HS. The HS does not rely on a stereotyped brain affect, we performed psychophysiologic interaction
organization, as is the case for well-defined states of analyses,11 looking for regions that would respond to
vigilance such as sleep stages.25,26 On the contrary, in noxious stimulations under the modulatory action of the
the HS, brain work may be directed at will to certain ACC specifically in the HS. No significant results were
tasks. In our case, perception of noxious stimulation was obtained by these analyses, possibly because of the small
at the center of subjects’ concern. Other cognitive tasks number of observations. Consequently, the physiologic
may be generated during the HS, such as memory recall significance of the midcingulate activation in the HS
and automatic writing. Each of these cerebral functions during noxious stimulation remains putative.
is likely to correspond to a different brain activation It is unlikely that opioid neurotransmission underlies
pattern in the HS. This suggestion is in good agreement the midcingulate activation we observed under the HS,
with the results of Grond et al.,27 showing that hypnot- although the ACC contains high concentrations of opi-
ically induced catalepsy was related to increased glucose oid receptors and peptides.28,29 Indeed, psychopharma-
metabolism in the sensorimotor cortex. cologic studies showed that hypnotic analgesia was not
altered by the administration of naloxone.30 Further-
State-by-stimulation Interaction: The Effect of the more, Adler et al.20 showed that fentanyl, an opioid
HS on Pain Perception agonist that has powerful analgesic properties, causes an
The results of the interaction analysis, especially using activation rather than a deactivation of midcingulate
a multiple regression approach, confirmed a differential cortex. In other words, under fentanyl administration,
modulation in midcingulate (ACC) activity in response to ACC blood flow increases while pain perception de-
noxious stimuli, in the specific context of HS, as com- creases, in contrast to what is observed in the HS.
pared with control states. The CBF in the ACC increases It is also unlikely that the ACC might modulate pain
steeply in relation to pain ratings, in the specific context perception during the HS through attentional mecha-
of the HS. Given our experimental setting, this result nisms. The midcingulate cortex that we show activated
would suggest that ACC activity plays a role in decreas- in our study has been related to pain perception,
ing pain ratings. whereas the more anterior portions of the ACC are
The mechanisms by which the midcingulate cortex involved in attention-demanding tasks.31,32 These ana-
may modulate response to noxious stimuli remain un- tomic considerations suggest that attentional processes

Anesthesiology, V 92, No 5, May 2000

1266

FAYMONVILLE ET AL.

ed cingulate areas, supplementary motor area) might

also allow the midcingulate cortex to organize the most
appropriate behavioral response, taking into account the
affective component of stimuli to the pain perception.

Comparison with the Data of Rainville et al.5

A recent PET study explored the neuroanatomic cor-
relates of “pain affect” during hypnosis.5 The investiga-
tors specifically used hypnotic suggestions to increase or
decrease noxious unpleasantness, seemingly without af-
fecting pain sensation by separating sensory and affec-
tive pain perception. It should be emphasized that these
behavioral results are in contrast to those of Kiernan et
al.,15 who showed that intensity and unpleasantness
remain highly correlated during the HS (r ⫽ 0.88). Nev-
ertheless, during HS, Rainville et al.5 observed significant
changes in pain-evoked activity within the ACC in the
HS, consistent with the encoding of perceived unpleas-
antness. In the authors’ view, this suggested “a specific
encoding for noxious unpleasantness in the ACC.” Our
results confirm that noxious unpleasantness during the
HS is related to ACC activity, in keeping with this previ-
ous PET study. Indeed, the coordinates of the ACC acti-
vation (coordinates: ⫺2, 18, 22 mm) are close to those of
Rainville et al.5 (coordinates: ⫺1, 25, 29 mm; distance in
y and z direction ⫽ 7 mm).
However, using our hypnotic technique, we were able
to show that the HS reduces both noxious perception
and unpleasantness. This effect is specific to the HS and
cannot be accounted for by the subject being distracted
Fig. 5. (A) Brain area in which blood flow increases in propor- from noxious stimuli: as a control, MI did not signifi-
tion to pain sensation ratings, in the specific context of the
hypnotic state: the ventral part of the midcingulate cortex (ten-
cantly decrease pain ratings. The decrease in both affec-
tatively area 24ⴕa). Results are displayed on the subject’s T1- tive and sensory aspects of pain perception is, of course,
weighted average magnetic resonance imaging scan, normal- critical for hypnosis that is used to reduce perioperative
ized to the same standardized space. (B) Plot of adjusted
anterior cingulate cortex blood flow versus pain perception
pain. Furthermore, in HS, the ACC responds to both
ratings. The results show that there is a significant difference perceptive and affective aspects of pain sensation.
(P ⴝ 0.047, voxel level) in pain ratings versus regional cerebral Consequently, our functional data extend the results of
blood flow regression slopes between the hypnotic state
(green) and control conditions (red).
Rainville et al.5 by showing that both affective and sen-
sory responses to noxious stimulation are reduced in the
specific context of HS, and this reduction is mediated by
were probably not responsible for the analgesia during
the ACC.
the HS.
In conclusion, pain perception by normal subjects can
From the anatomic standpoint, the ACC is anatomically
be modified by the HS. This modulatory effect of the HS
and functionally heterogeneous.33,34 Anatomically speak-
seems mediated by the midcingulate cortex activity. In-
ing, the midcingulate cortex is in critical position to receive
deed, the reduction of pain perception correlated with
both the sensory noxious aspects from the somatosen-
ACC activity specifically in the HS.
sory areas and insula, and the affective component of
noxious stimuli, encoded in amygdaloid complexes and The authors thank Professors R. S. J. Frackowiak and K. J. Friston
pregenual ACC.35 Functional relationships with nearby (Wellcome Department of Cognitive Neurology, Institute of Neurol-
premotor areas of the medial frontal cortex (motor-relat- ogy, London, United Kingdom) for kindly providing the statistical

Anesthesiology, V 92, No 5, May 2000

 1267

PET STUDY OF PAIN MODULATION BY THE HYPNOTIC STATE

parametric mapping software, and Mrs. C. Mesters, Mr. P. Hawotte, 19. Casey KL, Minoshima S, Berger KL, Koeppe RA, Morrow TJ, Frey KA:
and Mr. J-L. Génon for technical assistance. Positron emission tomographic analysis of cerebral structures activated spe-
cifically by repetitive noxious heat stimuli. J Neurophysiol 1994; 71:802–7
References 20. Adler LJ, Gyulai FE, Diehl DJ, Mintun MA, Winter PM, Firestone
LL: Regional brain activity changes associated with fentanyl analgesia
1. Faymonville ME, Fissette J, Mambourg P, Roediger L, Joris J, Lamy
elucidated by positron emission tomography [published erratum ap-
M: Hypnosis as an adjunct therapy in conscious sedation for plastic
surgery. Reg Anesth 1995; 20:145–51 pears in Anesth Analg 1997; 84:949]. Anesth Analg 1997; 84:120 – 6
2. Faymonville ME, Mambourg P, Joris J, Vrijens B, Fissette J, Albert 21. Derbyshire SW, Jones AK, Devani P, Friston KJ, Feinmann C,
A, Lamy M: Psychological approaches during conscious sedation. Hyp- Harris M, Pearce S, Watson JD, Frackowiak RS: Cerebral responses
nosis versus stress reducing strategies: A prospective randomized to pain in patients with atypical facial pain measured by positron
study. Pain 1997; 73:361–7 emission tomography. J Neurol Neurosurg Psychiatry 1994; 57:
3. Faymonville ME, Meurisse M, Fissette J: Hypnosedation, a valu- 1166 –72
able alternative to traditional anaesthetic techniques. Acta Chir Belg 22. Hsieh JC, Stahle-Backdahl M, Hagermark O, Stone-Elander S,
1999; 99:141– 6 Rosenquist G, Ingvar M: Traumatic nociceptive pain activates the
4. Meurisse M, Hamoir E, Defechereux T, Gollogly L, Postal A, Joris hypothalamus and the periaqueductal gray: A positron emission tomog-
J, Faymonville M: Bilateral neck exploration under hypnosedation: A raphy study. Pain 1996; 64:303–14
new standard of care in primary hyperparathyroidism? Ann Surg 1999; 23. Xu X, Fukuyama H, Yazawa S, Mima T, Hanakawa T, Magata Y,
229:401– 8 Kanda M, Fujiwara M, Shindo K, Nagamine T, Shibasaki H: Functional
5. Rainville P, Duncan GH, Price DD, Carrier B, Bushnell MC: Pain localization of pain perception in the human brain studied by PET.
affect encoded in human anterior cingulate but not somatosensory Neuroreport 1997; 8:555–9
cortex. Science 1997; 277:968 –71 24. Rainville P, Hofbauer RK, Paus T, Duncan GH, Bushnell MC,
6. Hilgard ER, Lauer LW, Morgan AH: Manual for Standard Profile Price DD: Cerebral mechanisms of hypnotic induction and suggestion.
Scales of Hypnotic Susceptibility, Forms I and II. Palo Alto, Consulting
J Cogn Neurosci 1999; 11:110 –25
Psychologists Press, 1963
25. Maquet P, Peters J, Aerts J, Del Fiore G, Degueldre C, Luxen A,
7. Frackowiak RSJ, Friston KJ, Frith CD, Dolan RJ, Mazziotta JC:
Franck G: Functional neuroanatomy of human rapid eye movement
Human Brain Function. San Diego, Academic Press, 1997
sleep and dreaming. Nature 1996; 383:163– 6
8. Friston K, Ashburner J, Frith C, Poline JB, Heather J, Frackowiak
RSJ: Spatial realignment and normalization of images. Hum Brain Mapp 26. Maquet P, Degueldre C, Del Fiore G, Aerts J, Peters J, Luxen A,
1995; 2:165– 89 Franck G: Functional neuroanatomy of human slow wave sleep. J Neu-
9. Talairach J, Tournoux P: Co-planar Stereotaxic Atlas of the Human rosci 1997; 17:2807–12
Brain. Stuttgart, George Thieme Verlag, 1988 27. Grond M, Pawlik G, Walter H, Lesch OM, Heiss WD: Hypnotic
10. Friston K, Holmes AP, Worsley KJ, Poline JB, Frith CD, Frack- catalepsy-induced changes of regional cerebral glucose metabolism.
owiak RSJ: Statistical parametric maps in functional imaging: A general Psychiatry Res 1995; 61:173–9
approach. Hum Brain Mapp 1995; 2:189 –210 28. Jones AK, Friston KJ, Qi LY, Harris M, Cunningham VJ, Jones T,
11. Friston KJ, Buechel C, Fink GR, Morris J, Rolls E, Dolan RJ: Feinman C, Frackowiak RS: Sites of action of morphine in the brain
Psychophysiological and modulatory interactions in neuroimaging. [letter]. Lancet 1991; 338:825
NeuroImage 1997; 6:218 –29 29. Pfeiffer A, Pasi A, Mehraein P, Herz A: Opiate binding sites in
12. Friston KJ, Frith CD, Liddle PF, Dolan RJ, Lammertsma AA, human brain. Brain Res 1982; 248:87–96
Frackowiak RSJ: The relationship between global and local changes in 30. Moret V, Forster A, Laverrière MC, Lambert H, Gaillard RC,
PET scans. J Cereb Blood Flow Metab 1990; 10:458 – 66 Bourgeois P, Haynal A, Gemperle M, Buscher E: Mechanism of analge-
13. Maquet P, Faymonville ME, Degueldre C, Delfiore G, Franck G, sia induced by hypnosis and acupuncture: Is there a difference? Pain
Luxen A, Lamy M: Functional neuroanatomy of hypnotic state. Biol 1991; 45:135– 40
Psychiatry 1999; 45:327–33 31. Davis KD, Taylor SJ, Crawley AP, Wood ML, Mikulis DJ: Func-
14. Plum F, Posner JB: The Diagnosis of Stupor and Coma, 3rd
tional MRI of pain- and attention-related activations on the human
Edition. Philadelphia, F.A. Davis Company, 1980
cingulate cortex. J Neurophysiol 1997; 77:3370 – 80
15. Kiernan BD, Dane JR, Phillips LH, Price DD: Hypnotic analgesia re-
32. Derbyshire SWG, Vogt BA, Jones AKP: Pain and Stroop interfer-
duces R-III nociceptive reflex: Further evidence concerning the multifactorial
ence tasks activate separate processing modules in anterior cingulate
nature of hypnotic analgesia [see comments]. Pain 1995; 60:39–47
16. Derbyshire SW, Jones AK, Gyulai F, Clark S, Townsend D, cortex. Exp Brain Res 1998; 118:52– 60
Firestone LL: Pain processing during three levels of noxious stimulation 33. Devinsky O, Morrell MJ, Vogt BA: Contributions of anterior
produces differential patterns of central activity. Pain 1997; 73:431– 45 cingulate cortex to behaviour. Brain 1995; 118:279 –306
17. Derbyshire SW, Jones AK: Cerebral responses to a continual 34. Vogt BA, Sikes RW, Vogt LJ: Anterior cingulate cortex and the
tonic pain stimulus measured using positron emission tomography. medial pain system, Neurobiology of Cingulate Cortex and the Limbic
Pain 1998; 76:127–35 Thalamus: A Comprehensive Treatise. Edited by Vogt BA, Gabriel M.
18. Jones AK, Brown WD, Friston KJ, Qi LY, Frackowiak RS: Cortical Boston, Birkhauser, 1993, pp 313– 44
and subcortical localization of response to pain in man using positron 35. Vogt BA, Pandya DN: Cingulate cortex of the Rhesus monkey: II.
emission tomography. Proc R Soc Lond B Biol Sci 1991; 244:39 – 44 Cortical afferents. J Comp Neurol 1987; 262:271– 89

Anesthesiology, V 92, No 5, May 2000