OAP

PetSdaacsie

okoabtra

D

A

1d

pioid and Placebonalgesia Share the Same Network

redrag Petrovic, MD, PhD

The relationship between the endogenous opioid system and placebo analgesia has fas-cinated the research community for decades. Using functional imaging methods, it is nowpossible to study the underlying opioid and placebo processes in the brain. Such studieshave shown overlapping activity in the anterior cingulate cortex (ACC) and the insulastretching into the orbitofrontal cortex. Because the ACC is involved in top-down atten-tional regulation, this region may mediate the interaction between higher cognitive pro-cesses and the endogenous opioid network. Moreover, data also indicate that the ACC mayexert its modulation through the brainstem opioid network. The orbitofrontal cortex alsoshows placebo-dependent activation, but it does not have similar access to the opioidnetwork. Thus, this region may be involved in other aspects of the placebo response suchas processing treatment expectations.Semin Pain Med 3:31-36 © 2005 Elsevier Inc. All rights reserved.

KEYWORDS placebo-analgesia, pain, opioid, positron emission tomography, functional mag-netic resonance imaging, anterior cingulate cortex, insula, orbitofrontal cortex, brainstem

ts

TOfatrofnwain

gcag

cep

lacebo analgesia research had the first major scientificbreakthrough when it was discovered that the placebo

ffect depended on the endogenous opioid system,1 a findinghat has been replicated in several studies (eg, refs. 2-4).ince this discovery, complex and well-designed behavioralesigns have shown strong relationships between placebonalgesia and treatment expectation,2,5-8 drug conditioning,2

utonomic correlates,4 and somatotopic dependence.3 Re-ently, a third important step has been taken in placebo re-earch, that is, unraveling the underlying neuronal correlatesnvolved in the placebo analgesia response using positronmission tomography (PET) or functional MRI (fMRI).9-12

Placebo analgesia is a unique phenomenon because it isne of few higher cognitive top-down processes in which anown specific neuromodulatory system operates, ie, thepioid system. However, the endogenous opioid system islso involved in a broad range of emotional processes shownoth in animals and in humans.13-15 This knowledge makeshe placebo effect interesting not only for understanding painegulation but emotional regulation in general. This reviewrticle will try to relate the described placebo network9-12 to

epartment of Clinical Neuroscience, Karolinska Institute, Stockholm, Swe-den.

ddress correspondence to Predrag Petrovic, MD, MPH, MRC, Dept ClinicalNeuroscience, Karolinska Institute/Karolinska Hospital, 171 76 Stock-

hholm, Sweden. E-mail: predrag.petrovic@cns.ki.se

537-5897/05/$-see front matter © 2005 Elsevier Inc. All rights reserved.oi:10.1016/j.spmd.2005.02.005

he endogenous opioid system in the cortex and the brain-tem.

he Opioid System in the Brainpioid receptors in the central nervous system (CNS) are

ound in the entire neuroaxis, eg, in the cortex, brainstem,nd spinal cord.16-21 Although these receptors are widespreadhroughout the CNS, the localization is not diffuse but highlyegional. It has been proposed that networks containing opi-id receptors may exert analgesic effects through several dif-erent mechanisms.22 These include modulation of spinaloxious input (in the dorsal horn and in the ascending path-ays), direct control of cortical and brainstem structures that

re involved in pain processing, or regulation of the ascend-ng forebrain systems. At present, the modulation of the spi-al cord has been best described.17,18

The brainstem opioid system consists of a network of re-ions, including the periaqueductal gray (PAG), the parabra-hial nucleus, and the rostral ventromedial medulla , whichre critically involved in descending opioid dependent anal-esia (for a complete review, see refs. 17,18).

The opioid receptor network is less characterized in theortex, especially in the more developed human brain. How-ver, autoradiographic studies of postmortem human andrimate brains and PET studies of opioid receptors in the

uman subjects have started to reveal a cortical opioid sys-

31

ttPiolr

prhtrrRrrc

(iwtaaprost

ttbsosi

CtTtHtc

beTccipfdo

32 P. Petrovic

em. The autoradiographic studies indicate high concentra-ions of opioid receptors not only in the brainstem, eg, theAG, the intralaminar, and medial thalamic nuclei, but also

n the cingulate cortex and prefrontal cortex.19-21 These andther animal studies have suggested that the anterior cingu-ate cortex (ACC) has one of the highest levels of opioideceptor bindings in the cortex.23

PET studies using radioactive opioid [11C]-diprenor-hine, which indicates the mu-, delta-, and kappa-opioideceptor availability, have confirmed previous animal anduman autoradiography findings.24-26 Although no quantita-ive analysis of the entire brain has been presented, the opioideceptor images from these PET studies suggest a high opioideceptor concentration in the insula and the frontal cortex.aw data indicate that the binding potential is highest in theostral parts of the anterior cingulate cortex (rACC).24-26 Theeceptor imaging studies have also indicated high opioid re-eptor binding potential in basal ganglia and thalamus.

A similar network has showed increased neuronal activityusing PET and measuring the regional blood flow as anndex of the underlying neuronal activity) during treatmentith opioid-receptor agonists, such as remifentanil and fen-

anyl.9,27-30 In 2 of the studies, it was shown that remifentanil,mu-opioid compound with a short half life, increases the

ctivity in several regions known to be involved in pain-rocessing and containing a high concentrations of opioideceptors (eg, Fig. 1).9,30 Opioid-dependent increases werebserved in the caudal and rostral ACC, midanterior insula,tretching into the orbitofrontal/temporopolar cortex, and in

he brainstem. Common for all previously presented func- p

ional imaging studies was an observed activity increase inhe rostral ACC in response to opioid treatment. Thus, it maye suggested that a relationship exists between the opioidystem and the rACC. In summary, autoradiographic studies,pioid-receptor imaging studies, and functional imagingtudies all indicate that a specific opioid-rich network existsn the cortex that includes the ACC and the anterior insula.

onditions That Activatehe Endogenous Opioid Systemhe complexity of the endogenous opioid network indicates

hat it is an important regulatory system for the organism.owever, it is less elaborated in which natural situations

hese systems are active and their role for increasing thehances of the survival of the organism.

In animal studies, contexts that induce fear and stress haveeen described as important for the activation of the endog-nous opioid system, ie, fear- or stress-related analgesia.31

his state has been induced either using noxious shocks,onditioning a noxious shock with an neutral stimulation (ie,onditioned stress induced analgesia) or setting up a contextn which fear is thought to be induced in the animal, eg,utting an animal in the same cage as its predator (see ref. 31or details). These triggers will activate an opioid system me-iated via the amygdala, which then activates the brainstempioid system via the PAG.31

It has been hypothesized that several contexts may be im-

Figure 1 Comparison betweenplacebo-induced activity andopioid-induced activity in thecortex. The short-lasting mu-selective opioid remifentanilactivated cortical areas such asthe ACC (a) and the anteriorinsula/temporopolar cortex(stretching into the Obfc) (b).However, the activity in theObfc was only observed in theconjunction between the insu-la/temporopolar cortex and theObfc, possibly representingonly smoothing effect of thedata, ie, the core region of theObfc was not activated byremifentanil (c). Placebo-de-pendent activity was observedin the rACC (d), the right ante-rior insula (e), and in an exten-sive part of the lateral Obfcdominated in the right hemi-sphere (f). (The image is pre-sented on an SPM-template atwww.fil.ion.ucl.ac.uk/spm.)(Color version of figure isavailable online.)

ortant for activating endogenous opioid systems for the in-

dvptvgo

mmsmstpwitv

PAdvbctgptegnp

cticatscatpfta

PaWta

trTcrfwrwst

ldasiasiaiswmstt

FIie(teis

Opioid and placebo analgesia share the same network 33

uction of analgesia in humans, although a great deal ofariability exists in these studies.17,32 The placebo response isrobably the best-described experimental situation in whichhe endogenous opioid system in humans is naturally acti-ated1-4 and, thus, research on the placebo response is ofeneral importance also for understanding the endogenouspioid system.Apart from the analgesic function the opioid subsystemsay be specifically involved in different motivational andood regulations. Emotional responses in rats have been

uggested to be dependent on the opioid system.13,31 Evenore intriguing, it has been suggested that the mu-opioid

ystem may be involved in emotional processing and regula-ion in humans.14,15 Thus, its role is not just to regulate painerception but to modulate the general state of the organism,hich also includes a change of the perceived pain. This role

mplies that the opioid effect can be understood only in rela-ion to the external context and the internal state (eg, moti-ation and emotion).

lacebo and Cognitionlthough opioid conditioning is one possible mechanism in-ucing analgesia,2 the placebo effect also must be accessibleia higher cognitive processes because it is clearly affected byeliefs, attitudes, and conscious expectations.32,33 Nonspe-ific stimuli in the context, which are coded in higher cogni-ive networks, may both induce and modulate placebo anal-esia. An impressing and well-studied higher cognitiverocess that influences placebo response is expectations of areatment outcome.2,5-8 Thus, the belief that a treatment isffective directly correlates with the degree of placebo anal-esia.6 Therefore the placebo effect must also contain a cog-itive process converting the expectation to a changed painrocessing induced by top-down modulation.Apart from containing a large concentration of opioid re-

eptors, the ACC is involved in higher cognitive attentionalasks. Modern theories of attention suggest that the ACC isnvolved in conflict monitoring or conflict resolution.34-36 Aonflict implies that two different processes compete for thettentional space in the brain and therefore must be con-rolled.37 The meta-analysis by Bush and coworkers35

howed that the ACC may be divided into a cognitive (mid-audal ACC) and an emotional region (rACC). Moreover, theuthors suggested that the rACC also is involved attentionalasks but in the emotional–motivational domain. Becauseain may be viewed as a homeostatic emotion38 and a there-ore a conflicting emotional process competing for the atten-ional space, it may be suggested that the rACC is involved inttentional processing on the pain experience.39

lacebo Analgesiand the Opioid Networke have suggested9 that the rACC is involved in the interac-

ion between attention and the opioid system in placebo an-

lgesia because of the dense concentration of opioid recep- p

ors found in the ACC23 and its involvement in tasksequiring conflict resolution (as mentioned previously).35,36

hus, rACC may be viewed as the region in which higherognitive attentional processes have access to a specific neu-omodulatory system. In line with this suggestion, we per-ormed a PET study in which an increased neuronal activityas observed during placebo analgesia in the same region of

ostral ACC as maximally activated when the same subjectsere treated with opioids9 (Fig. 1). Another activation ob-

erved in the placebo condition encompassed the orbitofron-al cortex (Obfc) stretching in to the anterior insula.

Placebo-dependent activation of the rACC9 has been rep-icated recently (Fig. 2).10-12 A similar ACC activation wasescribed by Bingel and coworkers12 in placebo-induced an-lgesia. In the article by Wager and coworkers,10 2 placebotudies were performed while the underlying neuronal activ-ty was studied using functional MRI in both the anticipationnd the induction phase of a painful stimulus. Both studieshowed placebo-dependent activations of the rACC as well asn the lateral Obfc in the anticipation phase. However, thesectivations were not observed during the pain stimulationtself. Instead, decreased activations in the ACC were ob-erved in the placebo pain phase. The authors suggest thathen there is a preparatory phase, the placebo-dependentodulatory effect may be induced before the actual painful

timulation, leading to a decreased pain processing duringhe stimulation. This is a fascinating suggestion, indicatinghat we are able to activate the endogenous system in a pre-

igure 2 Placebo-dependent increases and decreases in the ACC.ncreased activity is indicated with gray, and decreased activity isndicated with white circles. In the studies by Wager and cowork-rs,10 the activity increases were observed in the anticipation phasegray circles), whereas the decreases were observed in the stimula-ion phase (white circles). In the study by Lieberman and cowork-rs,11 only decreases were observed in the ACC (white circle). (Themage is presented on an SPM-template at www.fil.ion.ucl.ac.uk/pm.)

erceptive manner. The decreased activation observed some-

wp

pscppmahiuatitamtc

sdetrpp

dtodasmspvittposw

ieriso

IaIAsampottmtbcbsAoacsi

PaSbr

Factoslo

34 P. Petrovic

hat later would then indicate an attenuated processing ofain unpleasantness processed in the ACC.40

In the study by Lieberman and coworkers11 no increasedlacebo-dependent activation was observed in the rACC; in-tead, a negative correlation was observed between the pla-ebo degree and the activity in mid-caudal ACC (a moreosterior region of the ACC). However, in this study thelacebo effect was induced during a 2-week period and acuteodulatory effects were not studied. The 2 last described

rticles10,11 point to some of the difficulties in the study ofigher cognitive modulations of pain, that is, similar regions

n the ACC are involved in processing perception of painnpleasantness (possibly inducing relative decrease of ACCctivity during placebo treatment) as well as in pain regula-ion (possibly inducing relative increases of ACC activity dur-ng placebo treatment). Although the pain regulatory activa-ions seem to be somewhat more rostral39 the 2 areas ofctivity may partly overlap and the pain-related decreasesay attenuate placebo induced increase in activity. Never-

heless, 4 of the 5 presented placebo experiments9,10,12 indi-ate the involvement of the rACC in placebo effect (Fig. 2).

When it comes to the orbitofrontal cortex, 4 of the pre-ented functional imaging experiments indicate a placebo-ependent activation of this region9-11 (Fig. 3) with a prefer-nce for the right Obfc.9,11 Few functional imaging studies onhe opioid effect have described extensive or any opioid-elated activations of the Obfc, which may indicate that otherrocessing components of the placebo response are com-

igure 3 Placebo-dependent increases in the lateral Obfc. Increasedctivity is indicated with gray circles. In both studies by Wager andoworkers,10 the placebo activations of the Obfc were observed inhe anticipation phase. In all other studies, the Obfc activations werebserved in the stimulation phase. The effect was not significant intudy 2 by Wager and coworkers possibly because there were prob-ems with attenuation effects in this region. (The image is presentedn an SPM-template at www.fil.ion.ucl.ac.uk/spm.)

uted in this region. As an example, we observed placebo- c

ependent increases in the Obfc that were clearly more ex-ensive than the opioid-dependent increases (Fig. 1).9 Thepioid-induced Obfc activation was present only on the bor-er between the temporopolar/insular activations. In fact, thectivations observed in the Obfc may more represent amoothing effect of the data. A key function of the Obfc is toonitor and modulate the motivational value of external

timuli based on their coupling to primary re-enforcers41-44 toerform a goal-directed behavior.45 Because the Obfc is in-olved in attributing external stimuli a relative value depend-ng on the internal states and external contexts it is temptingo suggest that this region is involved in inducing the expec-ation of treatment and how the expectation should affect theain experience, which may be a process preceding the opi-id-dependent modulation of pain perception. Such a biasignal may be used by the ACC that is in a position to interactith the endogenous opioid system (Fig. 4).The placebo-induced activity in the Obfc also stretched

nto the temporopolar region/anterior insula, which is inter-sting because the insula has a high concentration of opioideceptors indicated by both receptor imaging and functionalmaging studies (as mentioned previously) and also has beenuggested to be involved in a meta-representation of the statef the body associated with emotional awareness.38

nteraction Between rACCnd Brainstem Opioid Systems

t has previously been proposed that the opioid system in theCC has access to the powerful opioid system in the brain-tem (Fig. 4).18,23 We suggested that this is one of the mech-nisms by which the higher cognitive systems in the ACCay influence nociceptive input to the brain and therebyain perception.9 We performed a regression analysis andbserved that both in placebo analgesia and opioid analgesiahe rACC activity correlated with the brainstem, a findinghat was not observed for the pain condition without treat-ent. Although no causality can be shown, the data indicate

hat such a mechanism may exist. A similar regression haseen replicated in the placebo analgesia study by Bingel andoworkers.12 Moreover, a similar functional connection haseen shown in pain-distraction46 and prolonged tonic pain,47

uggesting that other pain regulatory conditions may involveCC-mediated opioid-dependent control of the brainstempioid system. Wager and coworkers did not supply such annalysis with the rACC but showed that also the orbitofrontalortex has a similar functional connectivity with the brain-tem.10 This supports the conclusion that both regions workn the same modulatory network.

lacebo Respondersnd the Opioid System

everal studies have indicated that there is a large variabilityetween individual subjects in the placebo response. Theeason for this is probably multifactorial. One factor that may

rucially affect the placebo analgesic response is the under-

loioppmiiisdrosniclss

CRaccsasbfTn

bfhrnosrwMiot

R

PM-te

Opioid and placebo analgesia share the same network 35

ying opioid system, that is, it would not possible to inducepioid-dependent placebo analgesia in a subject who is lack-ng an endogenous opioid system. Of course, there are nopioid-knockout subjects to test whether they respond tolacebo or not. However, data indicate that the opioid systemossibly vary in different subjects.48 Although this findingay have different causes, genetic factors are probably highly

nvolved. In line with this suggestion one recent study hasndicated that the COMT gene is directly involved in regulat-ng the opioid system and the functionality of the opioidystem in human subjects.49 In our placebo study, we crudelyivided the subjects in placebo responders and placebo non-esponders. We then compared the activations elicited by thepioids (ie, remifentanil) and observed that the placebo re-ponders activated the rACC whereas the nonresponders didot. This analysis was powerful because we specifically stud-

ed the opioid response (and not placebo response) in pla-ebo responders versus. nonresponders. The results are inine with a more effective opioid system in the placebo re-ponders. However, this subject sample was very small andhould be repeated in a larger subject sample.

onclusionecent imaging studies have started to disclose a functional–natomical relationship between the opioid and placebo pro-esses in the human brain. An emerging view suggests that aomplex network in the brain is underlying the placebo re-ponse (Fig. 4), consisting of both opioid rich regions in ACCnd insula (involved in subprocesses of the placebo effectuch as attentional modulation and emotional awareness),ut also regions not activated by opioids such as the orbito-rontal cortex (involved in the processing of expectation).his network induces top-down control of a powerful opioid

Figure 4 The opioid network in the cortex and the brainan opioid-rich network in the brainstem, including the pthe rostral ventrolateral medulla (RVM). This networkamygdala in fear-induced analgesia. We suggest that thanterior insula and the Obfc, is similarly involved in branot highly activated by opioids, it may be involved prtreatment expectations. (The image is presented on an S

etwork in the brainstem. Of course, other regions also may 1

e involved in the placebo process, eg, the dorsolateral pre-rontal cortex is probably involved in short-term memoryolding the nonemotional context supporting the placeboesponse on-line. Another region possibly belonging to thisetwork is the amygdala because it is involved in a similarpioid-dependent control of the brainstem in the fear re-ponse. In a near future, displacement studies of the opioideceptor system will further show whether the opioid net-ork may be directly involved in the placebo response.oreover a combination between genetic and imaging stud-

es may indicate whether different genotypes affecting thepioid system also affects the opioid phenotype and therebyhe placebo response.

eferences1. Levine JD, Gordon NC, Fields HL: The mechanism of placebo analge-

sia. Lancet 2:654-657, 19782. Amanzio M, Benedetti F: Neuropharmacological dissection of placebo

analgesia: expectation-activated opioid systems versus conditioning-activated specific subsystems. J Neurosci 19:484-494, 1999

3. Benedetti F, Arduino C, Amanzio M: Somatotopic activation of opioidsystems by target-directed expectations of analgesia. J Neurosci 19:3639-3648, 1999

4. Pollo A, Vighetti S, Rainero I, Benedetti F: Placebo analgesia and theheart. Pain 102:125-133, 2003

5. Montgomery GH, Kirsch I: Classical conditioning and the placebo ef-fect. Pain 72:107-113, 1997

6. Price DD, Milling LS, Kirsch I, et al: An analysis of factors that contrib-ute to the magnitude of placebo analgesia in an experimental paradigm.Pain 83:147-156, 1999

7. Voudouris NJ, Peck CL, Coleman G: Conditioned response models ofplacebo phenomena: Further support. Pain 38:109-116, 1989

8. Voudouris NJ, Peck CL, Coleman G: The role of conditioning andverbal expectancy in the placebo response. Pain 43:121-128, 1990

9. Petrovic P, Kalso E, Petersson KM, Ingvar M: Placebo and opioid anal-gesia—imagine a shared neuronal network. Science 295:1737-1740,2002

d its interaction with the orbitofrontal cortex. There iseductal gray (PAG), the parabracial nucleus (PBN), ande controlled by the opioid-dependent processes in theas a part of an opioid-rich network also including theregulation in placebo analgesia. Given that the Obfc isg other components in the placebo response, such asmplate at www.fil.ion.ucl.ac.uk/spm.)

stem aneriaqumay be ACC,instemocessin

0. Wager TD, Rilling JK, Smith EE, et al: Placebo-induced changes in

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

3

3

3

4

4

4

4

4

4

4

4

4

4

36 P. Petrovic

FMRI in the anticipation and experience of pain. Science303:1162-1167, 2004

1. Lieberman MD, Jarcho JM, Berman S, et al: The neural correlates ofplacebo effects: A disruption account. Neuroimage 22:447-455, 2004

2. Bingel U, Lorenz J, Gläscher J, et al: Cognitive control of pain: A com-mon system for pain expectation and placebo-analgesia—A single trialfMRI study. Neuroimage 23:224-232, 2004

3. Filliol D, Ghozland S, Chluba J, et al: Mice deficient for delta- andmu-opioid receptors exhibit opposing alterations of emotional re-sponses. Nat Genet 25:195-200, 2000

4. Liberzon I, Zubieta JK, Fig LM, et al: mu-Opioid receptors and limbicresponses to aversive emotional stimuli. Proc Natl Acad Sci USA 99:7084-7089, 2002

5. Zubieta JK, Ketter TA, Bueller JA, et al: Regulation of human affectiveresponses by anterior cingulate and limbic mu-opioid neurotransmis-sion. Arch Gen Psychiatry 60:1145-1153, 2003

6. Atweh F, Kuhar MJ: Distribution and physiological significance of opi-oid receptors in the brain. Br Med Bull 39:47-52, 1983 [review]

7. Fields H, Basbaum A: Central nervous system mechanisms of painmodulation, in Wall P, Melzack R (eds): Textbook of Pain. New York,NY, Churchill Livingstone, 1999, pp 309-329

8. Fields H: State-dependent opioid control of pain. Nat Rev Neurosci5:565-575, 2004

9. Pfeiffer A, Pasi A, Mehraein P, Herz A: Opiate receptor binding sites inhuman brain. Brain Res 248:87-96, 1982

0. Sadzot B, Price J C, Mayberg HS, et al: Quantification of human opiatereceptor concentration and affinity using high and low specific activityand diprenorphine and positron emission tomography. J Cereb BloodFlow Metab 11:204-219, 1991 [erratum appears in J Cereb Blood FlowMetab 12:885, 1992]

1. Wamsley JK, Zarbin MA, Young WS, Kuhar MJ: Distribution of opiatereceptors in the monkey brain: an autoradiographic study. Neuro-science 7:595-613, 1982

2. Jensen TS: Opioids in the brain: Supraspinal mechanisms in pain con-trol Acta Anaesthesiol Scand 41:123-132, 1997 [review]

3. Vogt BA, Sikes RW, Vogt LJ: Anterior cingulate cortex and the medialpain system, in Vogt BA, Gabriel M (eds): Neurobiology of CingulateCortex and Limbic Thalamus: A Comprehensive Handbook. Boston,Birkhäuser, 1993, pp 313-344

4. Willoch F, Schindler F, Wester HJ, et al: Central poststroke pain andreduced opioid receptor binding within pain processing circuitries: A[11C]diprenorphine PET study. Pain 108:213-220, 2004

5. Willoch F, Tolle TR, Wester HJ, et al: Central pain after pontine infarc-tion is associated with changes in opioid receptor binding: A PET studywith 11C-diprenorphine. AJNR Am J Neuroradiol 20:686-690, 1999

6. Jones AK, Qi LY, Fujirawa T, et al: In vivo distribution of opioid recep-tors in man in relation to the cortical projections of the medial andlateral pain systems measured with positron emission tomography.Neurosci Lett 126:25-28, 1991

7. Adler LJ, Gyulai FE, Diehl DJ, et al: Regional brain activity changesassociated with fentanyl analgesia elucidated by positron emission to-mography. Anesth Analg 84:120-126, 1997 [published erratum ap-pears in Anesth Analg 84:949, 1997]

8. Casey KL, Svensson P, Morrow TJ, et al: Selective opiate modulation ofnociceptive processing in the human brain. J Neurophysiol 84:525-

533, 2000

9. Firestone LL, Gyulai F, Mintun M, et al: Human brain activity responseto fentanyl imaged by positron emission tomography. Anesth Analg82:1247-1251, 1996

0. Wagner KJ, Willoch F, Kochs EF, et al: Dose-dependent regional cere-bral blood flow changes during remifentanil infusion in humans. Apositron emission tomography study. Anesthesiology 94:732-739,2001

1. Fanselow MS: Neural organization of the defensive behavior systemresponsible for fear. Psychonomic Bull Rev 1:429-438, 1994

2. Price D: Placebo analgesia, in Price DD (ed): Psychological Mechanismsof Pain and Analgesia. Seattle, WA, IASP Press, 1999, pp 155-181

3. Wall P: The placebo and the placebo response, in Wall P, Melzack R(eds): Textbook of Pain. New York, Churchill Livingstone, 1999, pp1419-1430

4. Carter CS, Botvinick MM, Cohen JD: The contribution of the anteriorcingulate cortex to executive processes in cognition. Rev Neurosci 10:49-57, 1999 [review]

5. Bush G, Luu P, Posner MI: Cognitive and emotional influences inanterior cingulate cortex. Trends Cogn Sci 4:215-222, 2000

6. Paus T: Primate anterior cingulate cortex: Where motor control, driveand cognition interface. Nat Rev Neurosci 2:417-424, 2001

7. Stuss DT, Shallice T, Alexander MP, et al: A multidisciplinary approachto anterior attentional functions. Ann NY Acad Sci 769:191-211,1995[review]

8. Craig A: A new view of pain as a homeostatic emotion. Trends Neurosci26:303-307, 2003

9. Petrovic P, Ingvar M: Topical review: Imaging cognitive modulation ofpain processing. Pain 95:1-5, 2002

0. Rainville P, Duncan GH, Price DD, et al: Pain affect encoded in humananterior cingulate but not somatosensory cortex. Science 277:968-971,1997

1. Kringelbach ML, Rolls ET: Neural correlates of rapid reversal learningin a simple model of human social interaction. NeuroImage 20:1371-1383, 2003

2. O‘Doherty JP, Deichmann R, Critchley HD, et al: Neural responsesduring anticipation of a primary taste reward. Neuron 33:815-826,2002

3. Tremblay L, Schultz W: Relative reward preference in primate orbito-frontal cortex [see comments]. Nature 398:704-708, 1999

4. Gottfried JA, O’Doherty J, Dolan RJ: Encoding predictive reward valuein human amygdala and orbitofrontal cortex. Science 301:1104-1107,2003

5. Schoenbaum G, Setlow B: Integrating orbitofrontal cortex into prefron-tal theory: common processing themes across species and subdivisions.Learning Memory 8:134-147, 2001

6. Valet M, Sprenger T, Boecker H, et al: Distraction modulates connec-tivity of the cingulo-frontal cortex and the midbrain during pain—anfMRI analysis. Pain 109:399-408, 2004

7. Petrovic P, Petersson KM, Hansson P, et al: Brainstem involvement inthe initial response to pain. Neuroimage 22:995-1005, 2004

8. Zubieta J-K, Smith YR, Bueller JA, et al: Regional Mu opioid receptorregulation of sensory and affective dimensions of pain. Science 293:311-315, 2001

9. Zubieta JK, Heitzeg MM, Smith YR, et al: COMT val158met genotypeaffects mu-opioid neurotransmitter responses to a pain stressor. Sci-

ence 299:1240-1243, 2003