Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

INTRODUCTION

At first glance pain seems relatively straightforward ndashhitting onersquos thumb with a hammer hurts onersquosthumb Such experiences are easily understood with astructural-pathology model which supposes painprovides an accurate indication of the state of the tis-sues However on closer inspection pain is lessstraightforward Much of the pain we see clinicallyfits into this less straightforward category where paincannot be understood as a marker of the state of thetissues This paper argues that the biology of pain isnever really straightforward even when it appears tobe There are four key points (i) that pain does notprovide a measure of the state of the tissues (ii) thatpain is modulated by many factors from acrosssomatic psychological and social domains (iii) thatthe relationship between pain and the state of the tis-sues becomes less predictable as pain persists and (iv)that pain can be conceptualised as a conscious corre-late of the implicit perception that tissue is in dangerThese points will be discussed in light of their clinical

implications and will form the basis of one approachto understanding complex regional pain syndrome

Pain does not provide a measure of the state of thetissues

In 1965 the gate control theory2 was proposed toexplain the variable response of animals to noxiousstimuli The theory proposed that noxious input wasmodulated at the spinal cord by other non-noxiousinput from the periphery and by descending inputfrom higher centres That theory was interrogated inmany animal experiments (see Wall and McMahon3

for a review) A typical experiment would involve theinsertion of recording electrodes into the nociceptorsof the study animal applying a defined injury andrecording nociceptor activity Finally experimenterswould record behaviours of the animal that impliedthat the animal was in pain These behaviours mightbe relatively simple for example the reaction time of awithdrawal reflex They might be relatively complex

copy W S Maney amp Son Ltd 2007 DOI 101179108331907X223010

RECONCEPTUALISING PAIN ACCORDING TO MODERN PAIN SCIENCE

G LORIMER MOSELEY

Oxford Centre for fMRI of the Brain Department of Physiology Anatomy amp Genetics Oxford University Oxford UK

This paper argues that the biology of pain is never really straightforward even when it appearsto be It is proposed that understanding what is currently known about the biology of painrequires a reconceptualisation of what pain actually is and how it serves our livelihood Thereare four key points (i) that pain does not provide a measure of the state of the tissues (ii) thatpain is modulated by many factors from across somatic psychological and social domains (iii)that the relationship between pain and the state of the tissues becomes less predictable as painpersists and (iv) that pain can be conceptualised as a conscious correlate of the implicitperception that tissue is in danger These issues raise conceptual and clinical implicationswhich are discussed with particular relevance to persistent pain Finally this conceptualisationis used as a framework for one approach to understanding complex regional pain syndrome

Keywords Pain re-evaluation persistent pain pain syndrome

Physical Therapy Reviews 2007 12 169ndash178

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

for example the ratio between time spent in a non-preferred environment (eg illuminated box) with acool floor and time spent in a preferred environment(eg dark box) with a heated floor4

Two findings consistently emerged from those stud-ies First the injury or noxious stimulation initiatesthe change in behaviour Second neither pain behav-iour nor nociceptor activity hold an isomorphic rela-tionship with the state of the tissues By clearlydemonstrating these things those studies providedthe first experimental evidence that pain does not pro-vide a measure of the state of the tissues

One limitation of animal experiments is that theydo not tell us about pain Human experiments how-ever can Although it is difficult to justify injuringhuman volunteers it is possible to deliver non-harm-ful noxious stimuli for example brief thermal electri-cal or mechanical stimuli (see Handwerker andKobal5 for a review of various methods of experimen-tally inducing pain) By recording activity in nocicep-tors while simultaneously recording subjectsrsquo painratings experimenters have been able to evaluate therelationship between the state of the tissues (in theabsence of tissue damage) activity in nociceptorsand pain6

Human pain experiments corroborated both find-ings from the animal data Specifically noxious stim-ulation is necessary for nociceptor activity whichusually reflects the intensity of the stimulus and noci-ceptor activation does not provide an accurate mea-sure of the state of the tissues6 The humanexperiments went further because they showed thatthe relationship between pain ratings and nociceptoractivation is variable In fact some authors have pro-posed that the notion of nociceptors is misleadingbecause small diameter fibres (Aδ and C fibres)respond to very small (non-harmful) changes in theinternal state of the body7 That said some smalldiameter fibres are not responsive to small changes(so-called high-threshold neurons) and this sub-classof small diameter fibres may reflect what we call noci-ceptors Regardless it is clear that experimental stud-ies do not show an isomorphic relationship betweenpain and nociceptor activity nor between pain andthe state of the tissues Rather they show a variablerelationship that is modulated by many factors

Pain is modulated by many factors from across somaticpsychological and social domains

Anecdotal evidence that somatic psychological andsocial factors modulate pain is substantial ndash sport-related and war-related stories are common (seeButler and Moseley8 for several examples) However

numerous experimental findings corroborate theanecdotal evidence (see Fields et al9 for a review ofcentral nervous system mechanisms of modulation)Other factors that are known to modulate the painevoked by a standardised stimulus include inflamma-tory mediators (increase nociceptor activity) tissuetemperature (increased temperature increases noci-ceptor activity via summation) and blood flow(decreased blood flow increases nociceptor activityvia summation induced by H+ ions) See Meyer et al6

for a review of peripheral mechanisms of modulationExperiments that manipulate the psychological

context of a noxious stimulus often demonstrate cleareffects on pain although the direction of these effectsis not always consistent For example a large amountof literature concerns the effect of attention on painand of pain on attention10ndash22 Despite the wealth ofdata consensus is lacking some data suggest thatattending to pain amplifies it and attending awayfrom pain nullifies it but others suggest the opposite

Anxiety also seems to have variable effects on painSome reports link increased anxiety to increased painduring clinical procedures23ndash26 and during experimentallyinduced pain27 but other reports suggest no effect2829

Relevant reviews conclude that the influence of anxietyon pain is probably largely dependent on attention2830

Expectation also seems to have variable effects onpain As a general rule expectation of a noxious stimu-lus increases pain if the cue signals a more intense ormore damaging stimulus2231ndash35 and decreases pain if thecue signals a less intense or less damaging stimulus (seeFields34 and Wager36 for reviews) Further cues that sig-nal an impending decrease in pain for example theprocess of taking an analgesic usually decrease painThus expectation is thought to play a major role inplacebo analgesia3738

The common denominator of the effect of attentionanxiety and expectation on pain seems to be the underly-ing evaluative context or meaning of the pain That isdemonstrated by the consistent effect that some cognitivestates seem to have on pain For example catastrophicinterpretations of pain are associated with higher pain rat-ings in both clinical and experimental studies (see Sullivanet al39 for a review) Believing pain to be an accurate indi-cator of the state of the tissues is associated with higherpain ratings40 whereas believing that the nervous systemamplifies noxious input in chronic pain states increasespain threshold during straight leg raise41

The social context of a noxious stimulus also affects thepain it evokes Initiation practices and sadomasochisticsexual practices are two examples that highlight theimportance of social context Overall the effects of socialcontext are again variable but again seem to be under-pinned by the underlying evaluative context or meaning(see Butler and Moseley8 for a review of pain-related data

170 MOSELEY

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

and Moerman42 for exhaustive coverage of the role ofmeaning in health and medicine)

To review the very large amount of literature onsomatic psychological and social influences on pain isbeyond the scope of this paper However it is appropriateand clinically meaningful to reiterate the theme thatemerges from that literature that the influences are vari-able and seem to depend on the evaluative context of thenoxious input

The relationship between pain and the state of thetissues becomes weaker as pain persists

The nervous system is dynamic This means that the func-tional properties of individual neurones and of synergiesof neurones change in response to activity To review allthe changes that have been identified is beyond the scopeof this paper and the expertise of this author Howeverthe nature of the changes can be summarised thus thatthe neurones that transmit nociceptive input to the brainbecome sensitised as nociception persists and that thenetworks of neurons within the brain that evoke painbecome sensitised as pain persists The molecular and sys-tems biology of these changes have been discussed at sev-eral levels84344 The clinical manifestations of thesechanges are hyperalgesia (formerly painful stimuli

become more painful) and allodynia (formerly non-painful stimuli become painful) These terms are usedwidely most often in reference to tactile stimuli but alsoin reference to movement and to thermal stimuli

One aspect of the changes that occur when pain persistsis that the proprioceptive representation of the painfulbody part in primary sensory cortex changes45ndash47 Thismay have implications for motor control because theserepresentations are the maps that the brain uses to planand execute movement48 If the map of a body partbecomes inaccurate then motor control may be compro-mised ndash it is known that experimental disruption of corti-cal proprioceptive maps disrupts motor planning49 Thenotion of distorted proprioceptive representation hasbeen discussed with regard to its impact on motor con-trol5051 and more recently in a theoretical way withregard to pain52 Although exceptions exist53 there ismounting evidence that changes in cortical representationoccur in association with chronic pain and it is feasiblethat these changes may become part of the problem46

Conceptualising pain as a conscious correlate of theimplicit perception that tissue is in danger

The biology of pain is complex One response to thiscomplexity is to develop clinically viable conceptual

RECONCEPTUALISING PAIN ACCORDING TO MODERN PAIN SCIENCE 171

Fig 1 Many inputs affect the implicit perception of threat to body tissues labelled here as lsquoHow dangerous is this reallyrsquoThose inputs have wider meaning effects which in turn seems to affect anxiety attention and expectation The implicitperception of threat to body tissues determines pain and in turn influences other inputs

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

paradigms that incorporate what is now known aboutthat complexity One such paradigm that is gaining sup-port is the neuromatrix theory (see Melzack55 for a con-textual review)54 which conceptualises pain as oneoutput of the central nervous system that occurs whenthe organism perceives tissue to be under threat Thereare two important components of this conceptualisa-tion First there are other central nervous system out-puts that occur when tissue is perceived to be underthreat and second that it is the implicit perception ofthreat that determines the outputs not the state of thetissues nor the actual threat to the tissues (Fig 1)

When tissue is under threat a range of local and seg-mental responses occur For example inflammatorymediators are released the body part is usually with-drawn via short and long latency reflex loops there arerapid changes in blood flow and in the excitability ofperipheral nociceptors (so-called peripheral sensitisa-tion)56 The nociceptive system transforms this threatinto electrical activity in peripheral neurones If thismessage of threat is then transmitted by spinal neuronesto higher centres the responses become more complexFor example immune mediators are released into theblood stream57 voluntary and postural muscle activityare altered58 and conscious knowledge of the threat (iepain) may emerge Within this context pain will notemerge until the nociceptive input to the brain has beenevaluated albeit at an unconscious level (see Moseley59

and Gifford et al60 for further discussion)The second important component of the neuromatrix

theory is that pain depends on the perceived degree ofthreat This means that pain can be conceptualised as theconscious correlate of the implicit perception of threat tobody tissues859 That psychosocial factors are very impor-tant in most chronic pain states is well established61ndash65

This paper argues that the mass of data regarding psy-chosocial factors can be gathered within the proposedconceptualisation that pain is one output of the centralnervous system that occurs when the organism perceivestissue to be under threat The conceptualisation has limi-tations and strengths One limitation is that it does notattempt to describe the biology of implicit evaluation ofthreat nor of how this might emerge into consciousnessIn this sense it adds little to theories first proposeddecades ago (see for example Hebb66) However astrength of this conceptualisation is that it can easily beintegrated into a clinical context where making sense ofthe influence of factors from across somatic psychologi-cal and social domains is valuable

Implications for clinical practice

That pain does not reflect the state of the tissues butrather is a conscious driver of behaviour aimed at

protecting those tissues has implications for clinicalpractice One implication is that to base clinical rea-soning on what is currently known about the biologyof pain requires that the skills and knowledge of theclinician are broader than those related to anatomyand biomechanics That is the clinician must have asound knowledge of diagnostic tools tissue dynamicshealing and remodelling peripheral and central sensi-tisation and psychological and social factors thatmight affect the implicit perception of threat to bodytissues This information is readily available and thereis evidence that clinicians can understand modernconcepts with relatively limited training67 That saidit may be unrealistic to expect clinicians to keep up-to-date with progress in knowledge across these areasThis points to a strength of the conceptualisation ofpain as the conscious correlate of the implicit percep-tion threat to body tissues because the clinician canuse the conceptual model to guide treatment That israther than know and understand all the evidenceabout which somatic psychological and social factorshave been demonstrated to modulate pain and thenature of their modulation the clinician can considereach factor in terms of what effect it might have onthe implicit perception of threat This conceptualmodel seeks to synthesize that wide body of evidenceinto a principle

Another implication that is worthy of special men-tion is that patients should be helped to base their rea-soning about their condition and their pain onsimilar information This is important because teach-ing patients about modern pain biology leads toaltered beliefs and attitudes about pain40 andincreased pain thresholds during relevant tasks41

Moreover when education about pain biology isincorporated into physiotherapy management ofpatients with chronic pain pain and disability arereduced6869 A key objective of such education is toencourage patients to apply the same principle as thatadvocated for clinicians summarised here as lsquowhateffect might this (factor) have on the implicit percep-tion of threatrsquo or in patient-appropriate languagelsquohow does this affect the answer to the question howdangerous is this reallyrsquo8

USING THIS CONCEPTUALISATION TOUNDERSTAND CRPS AND GUIDE NEW

OPTIONS FOR MANAGEMENT

Complex regional pain syndrome (CRPS) is a debili-tating condition that can occur after minor traumaand sometimes without peripheral trauma for exam-ple post-stroke70 Much is known about the patho-physiology of CRPS including facilitated neurogenic

172 MOSELEY

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

inflammation7172 and tissue hypoxia73 at the injurysite7475 autonomic76 immune77ndash79 motor8081 tactile82ndash85

and proprioceptive86 dysfunction (Fig 2)The syndromic pattern of signs and symptoms

includes pain hyperalgesia allodynia dystoniaswelling abnormal blood flow abnormal sweatinghair and nail growth The sensitivity to provocationcan be remarkable for example elicitation of painswelling and (anecdotally) blood flow changes inresponse to imagined movements87 or when thepatient receives visual input that the limb is beingtouched even though it is not in fact being touched(lsquodysynchiriarsquo)88 The wide-spread and multisystemicnature of the pathophysiology of CRPS implies that

although CRPS is usually initiated by peripheralinsult it is a disorder of the central nervous system75

When one tries to make sense of such a multisystemicand exaggerated response to minor injury the conceptual-isation that pain is a conscious correlate of the implicitperception of the threat to body tissue can be useful Thatpain is just one output by which the brain might try toprotect the tissues ndash one aspect of a homeostaticresponse89 ndash lends itself to CRPS because the otherresponses are so patent That pain is a correlate of implic-itly perceived threat to body tissue rather than the state ofthe tissues or the actual threat to the tissues is particu-larly relevant to CRPS in the absence of any tissue orneural injury for example as a stress response90

RECONCEPTUALISING PAIN ACCORDING TO MODERN PAIN SCIENCE 173

Fig 2 Schematic overview of pathophysiology of complex regional pain syndrome Adapted from Janig and Baron75

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

Each of the pathological findings that have beendocumented in patients with CRPS might be consid-ered a protective response whether it be an immunemotor sensory vascular autonomic or consciousresponse consistent with attempts to protect the partin question by utilising immune motor sensory vas-

cular and autonomic systems as well as conscious-ness Reducing the threshold for activation of theseprotective responses would seem a particularly effec-tive way to protect the body part in question forexample making it so sensitive that even looking at itbeing touched activates a protective response88

174 MOSELEY

Fig 3 Response to different components of motor imagery Pain (A) and functional capacity (B) Recog = laterality recognitionwhereby patients make leftright judgements of pictured hands Imag = imagined movements Mirror = mirror movementsThree groups are shown group 1 undertook the motor imagery program group 2 performed imagined movements first thenlaterality recognition and then mirror movements group 3 performed laterality recognition followed by mirror movements andthen back to laterality recognition Note that group 1 had a largest reduction in pain Note also variable responses to imaginedand mirror movements depending on the order of components From Moseley1 and Moseley93 with permission

Recogniti

on

Neuropathic

pain scale

Task-specific

NRS

Time (weeks)

Time (weeks)

Imag

ined

Mirr

or

A

B

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

The challenge for those trying to understand CRPSaccording to this paradigm is to identify why theimplicit perception of threat to body tissues is soexaggerated in some patients and in some situationsbut not in others Fundamental to the paradigm isthat anything that modulates implicitly perceivedthreat should be relevant That means that psychoso-cial factors including anxiety depression attitudesand beliefs social context or work status may all playan important role Although patients with CRPS donot demonstrate a lsquotypicalrsquo psychosocial profile psy-chosocial contributors are probably relevant in themajority of cases Finally there is initial evidence fora genetic contribution to CRPS91 but more data arerequired to clarify that possibility

Clinical response to CRPS according to this paradigm

If CRPS is an exaggerated protective response then itseems sensible to devise treatment that aims first tofind a baseline that is sufficiently conservative to notelicit the unwanted protective responses (to lsquoget underthe radarrsquo) and second to expose the limb graduallyto threat while continuing to avoid elicitation of theunwanted responses This approach underpins gradedmotor imagery for CRPS9293 whereby patients begintraining by making leftright judgements of picturedlimbs It is known that this task activates cortical net-works that involve representation of the limb andpreparation for movement94 but this task does notactivate primary sensory and motor cortices95

Graded motor imagery progresses from leftright lat-erality judgements to imagined movements which doactivate primary sensory and motor cortices9596 andthen to mirror movements The order of these compo-nents seems to be important in the effect on pain anddisability in patients with chronic CRPS (Fig 3)1 Inpatients with acute (or anecdotally less severe) CRPSit may be sufficient to begin training (conceptualisedhere as exposure to threat) with mirror movements97

One of the key issues outlined earlier is that the ner-vous system changes when nociception and pain persistThere is a large amount of evidence that the cortical rep-resentation of the affected limb undergoes substantialchanges in patients with CRPS4774758398ndash100 and thesechanges have been implicated in the maintenance ofpathological pain syndromes (although see Moseley53

for a word of caution)46 If distorted cortical representa-tion contributes to CRPS then it would seem sensible toattempt to normalise cortical representation of the limbThis has been done in patients with phantom limbpain101 which is associated with changes in primary sen-sory cortex that are probably similar to those observedin CRPS (see Acerra et al102 for a review of common

findings in phantom limb pain stroke and CRPS) Inthat study with amputees sensory discrimination train-ing evoked normalisation of cortical representationimprovement in tactile acuity on the stump and reduc-tionelimination of phantom limb pain101 Increase intactile acuity normalisation of cortical representationand reduction in pain were positively related

Finally if CRPS reflects an exaggerated implicitperception of threat to body tissue then it wouldseem sensible to attempt to reduce the perception ofthreat One approach that has been studied exten-sively in other populations is the explanation to thepatient of the underlying biology of their painPreliminary data from patients with CRPS appearpromising103 but clinical trials are required

CONCLUSIONS

Extensive experimental data corroborate anecdotalevidence that pain does not provide a measure of thestate of the tissues and that pain is modulated bymany factors from across somatic psychological andsocial domains It is now known that as nociceptionand pain persist the neuronal mechanisms involvedin both become more sensitive which means that therelationship between pain and the state of the tissuesbecomes weaker and less predictable One paradigmwhich considers the current thought in pain biologyconceptualises pain as the conscious correlate of theimplicit perception of threat to body tissue This con-ceptualisation can be applied clinically to identify fac-tors from across somatic psychological and socialdomains that may affect the perceived threat to tissuedamage Further it suggests approaches to treatmentthat target those factors Evidence from clinical trialssuggests that clinical strategies based on this concep-tualisation can be effective in patients with disablingcomplex and chronic pain

ACKNOWLEDGEMENTS

The author is supported by the Nuffield DominionsTrust Oxford University UK The conceptual approachfor CRPS draws heavily on unpublished work of MichaelThacker Kingrsquos College London London UK

REFERENCES

1 Moseley GL Is successful rehabilitation of complex regionalpain syndrome due to sustained attention to the affected limbA randomised clinical trial Pain 200511454ndash61

2 Melzack R Wall PD Pain mechanisms a new theory Science1965150971ndash9

RECONCEPTUALISING PAIN ACCORDING TO MODERN PAIN SCIENCE 175

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

for example the ratio between time spent in a non-preferred environment (eg illuminated box) with acool floor and time spent in a preferred environment(eg dark box) with a heated floor4

Two findings consistently emerged from those stud-ies First the injury or noxious stimulation initiatesthe change in behaviour Second neither pain behav-iour nor nociceptor activity hold an isomorphic rela-tionship with the state of the tissues By clearlydemonstrating these things those studies providedthe first experimental evidence that pain does not pro-vide a measure of the state of the tissues

One limitation of animal experiments is that theydo not tell us about pain Human experiments how-ever can Although it is difficult to justify injuringhuman volunteers it is possible to deliver non-harm-ful noxious stimuli for example brief thermal electri-cal or mechanical stimuli (see Handwerker andKobal5 for a review of various methods of experimen-tally inducing pain) By recording activity in nocicep-tors while simultaneously recording subjectsrsquo painratings experimenters have been able to evaluate therelationship between the state of the tissues (in theabsence of tissue damage) activity in nociceptorsand pain6

Human pain experiments corroborated both find-ings from the animal data Specifically noxious stim-ulation is necessary for nociceptor activity whichusually reflects the intensity of the stimulus and noci-ceptor activation does not provide an accurate mea-sure of the state of the tissues6 The humanexperiments went further because they showed thatthe relationship between pain ratings and nociceptoractivation is variable In fact some authors have pro-posed that the notion of nociceptors is misleadingbecause small diameter fibres (Aδ and C fibres)respond to very small (non-harmful) changes in theinternal state of the body7 That said some smalldiameter fibres are not responsive to small changes(so-called high-threshold neurons) and this sub-classof small diameter fibres may reflect what we call noci-ceptors Regardless it is clear that experimental stud-ies do not show an isomorphic relationship betweenpain and nociceptor activity nor between pain andthe state of the tissues Rather they show a variablerelationship that is modulated by many factors

Pain is modulated by many factors from across somaticpsychological and social domains

Anecdotal evidence that somatic psychological andsocial factors modulate pain is substantial ndash sport-related and war-related stories are common (seeButler and Moseley8 for several examples) However

numerous experimental findings corroborate theanecdotal evidence (see Fields et al9 for a review ofcentral nervous system mechanisms of modulation)Other factors that are known to modulate the painevoked by a standardised stimulus include inflamma-tory mediators (increase nociceptor activity) tissuetemperature (increased temperature increases noci-ceptor activity via summation) and blood flow(decreased blood flow increases nociceptor activityvia summation induced by H+ ions) See Meyer et al6

for a review of peripheral mechanisms of modulationExperiments that manipulate the psychological

context of a noxious stimulus often demonstrate cleareffects on pain although the direction of these effectsis not always consistent For example a large amountof literature concerns the effect of attention on painand of pain on attention10ndash22 Despite the wealth ofdata consensus is lacking some data suggest thatattending to pain amplifies it and attending awayfrom pain nullifies it but others suggest the opposite

Anxiety also seems to have variable effects on painSome reports link increased anxiety to increased painduring clinical procedures23ndash26 and during experimentallyinduced pain27 but other reports suggest no effect2829

Relevant reviews conclude that the influence of anxietyon pain is probably largely dependent on attention2830

Expectation also seems to have variable effects onpain As a general rule expectation of a noxious stimu-lus increases pain if the cue signals a more intense ormore damaging stimulus2231ndash35 and decreases pain if thecue signals a less intense or less damaging stimulus (seeFields34 and Wager36 for reviews) Further cues that sig-nal an impending decrease in pain for example theprocess of taking an analgesic usually decrease painThus expectation is thought to play a major role inplacebo analgesia3738

The common denominator of the effect of attentionanxiety and expectation on pain seems to be the underly-ing evaluative context or meaning of the pain That isdemonstrated by the consistent effect that some cognitivestates seem to have on pain For example catastrophicinterpretations of pain are associated with higher pain rat-ings in both clinical and experimental studies (see Sullivanet al39 for a review) Believing pain to be an accurate indi-cator of the state of the tissues is associated with higherpain ratings40 whereas believing that the nervous systemamplifies noxious input in chronic pain states increasespain threshold during straight leg raise41

The social context of a noxious stimulus also affects thepain it evokes Initiation practices and sadomasochisticsexual practices are two examples that highlight theimportance of social context Overall the effects of socialcontext are again variable but again seem to be under-pinned by the underlying evaluative context or meaning(see Butler and Moseley8 for a review of pain-related data

170 MOSELEY

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

and Moerman42 for exhaustive coverage of the role ofmeaning in health and medicine)

To review the very large amount of literature onsomatic psychological and social influences on pain isbeyond the scope of this paper However it is appropriateand clinically meaningful to reiterate the theme thatemerges from that literature that the influences are vari-able and seem to depend on the evaluative context of thenoxious input

The relationship between pain and the state of thetissues becomes weaker as pain persists

The nervous system is dynamic This means that the func-tional properties of individual neurones and of synergiesof neurones change in response to activity To review allthe changes that have been identified is beyond the scopeof this paper and the expertise of this author Howeverthe nature of the changes can be summarised thus thatthe neurones that transmit nociceptive input to the brainbecome sensitised as nociception persists and that thenetworks of neurons within the brain that evoke painbecome sensitised as pain persists The molecular and sys-tems biology of these changes have been discussed at sev-eral levels84344 The clinical manifestations of thesechanges are hyperalgesia (formerly painful stimuli

become more painful) and allodynia (formerly non-painful stimuli become painful) These terms are usedwidely most often in reference to tactile stimuli but alsoin reference to movement and to thermal stimuli

One aspect of the changes that occur when pain persistsis that the proprioceptive representation of the painfulbody part in primary sensory cortex changes45ndash47 Thismay have implications for motor control because theserepresentations are the maps that the brain uses to planand execute movement48 If the map of a body partbecomes inaccurate then motor control may be compro-mised ndash it is known that experimental disruption of corti-cal proprioceptive maps disrupts motor planning49 Thenotion of distorted proprioceptive representation hasbeen discussed with regard to its impact on motor con-trol5051 and more recently in a theoretical way withregard to pain52 Although exceptions exist53 there ismounting evidence that changes in cortical representationoccur in association with chronic pain and it is feasiblethat these changes may become part of the problem46

Conceptualising pain as a conscious correlate of theimplicit perception that tissue is in danger

The biology of pain is complex One response to thiscomplexity is to develop clinically viable conceptual

RECONCEPTUALISING PAIN ACCORDING TO MODERN PAIN SCIENCE 171

Fig 1 Many inputs affect the implicit perception of threat to body tissues labelled here as lsquoHow dangerous is this reallyrsquoThose inputs have wider meaning effects which in turn seems to affect anxiety attention and expectation The implicitperception of threat to body tissues determines pain and in turn influences other inputs

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

paradigms that incorporate what is now known aboutthat complexity One such paradigm that is gaining sup-port is the neuromatrix theory (see Melzack55 for a con-textual review)54 which conceptualises pain as oneoutput of the central nervous system that occurs whenthe organism perceives tissue to be under threat Thereare two important components of this conceptualisa-tion First there are other central nervous system out-puts that occur when tissue is perceived to be underthreat and second that it is the implicit perception ofthreat that determines the outputs not the state of thetissues nor the actual threat to the tissues (Fig 1)

When tissue is under threat a range of local and seg-mental responses occur For example inflammatorymediators are released the body part is usually with-drawn via short and long latency reflex loops there arerapid changes in blood flow and in the excitability ofperipheral nociceptors (so-called peripheral sensitisa-tion)56 The nociceptive system transforms this threatinto electrical activity in peripheral neurones If thismessage of threat is then transmitted by spinal neuronesto higher centres the responses become more complexFor example immune mediators are released into theblood stream57 voluntary and postural muscle activityare altered58 and conscious knowledge of the threat (iepain) may emerge Within this context pain will notemerge until the nociceptive input to the brain has beenevaluated albeit at an unconscious level (see Moseley59

and Gifford et al60 for further discussion)The second important component of the neuromatrix

theory is that pain depends on the perceived degree ofthreat This means that pain can be conceptualised as theconscious correlate of the implicit perception of threat tobody tissues859 That psychosocial factors are very impor-tant in most chronic pain states is well established61ndash65

This paper argues that the mass of data regarding psy-chosocial factors can be gathered within the proposedconceptualisation that pain is one output of the centralnervous system that occurs when the organism perceivestissue to be under threat The conceptualisation has limi-tations and strengths One limitation is that it does notattempt to describe the biology of implicit evaluation ofthreat nor of how this might emerge into consciousnessIn this sense it adds little to theories first proposeddecades ago (see for example Hebb66) However astrength of this conceptualisation is that it can easily beintegrated into a clinical context where making sense ofthe influence of factors from across somatic psychologi-cal and social domains is valuable

Implications for clinical practice

That pain does not reflect the state of the tissues butrather is a conscious driver of behaviour aimed at

protecting those tissues has implications for clinicalpractice One implication is that to base clinical rea-soning on what is currently known about the biologyof pain requires that the skills and knowledge of theclinician are broader than those related to anatomyand biomechanics That is the clinician must have asound knowledge of diagnostic tools tissue dynamicshealing and remodelling peripheral and central sensi-tisation and psychological and social factors thatmight affect the implicit perception of threat to bodytissues This information is readily available and thereis evidence that clinicians can understand modernconcepts with relatively limited training67 That saidit may be unrealistic to expect clinicians to keep up-to-date with progress in knowledge across these areasThis points to a strength of the conceptualisation ofpain as the conscious correlate of the implicit percep-tion threat to body tissues because the clinician canuse the conceptual model to guide treatment That israther than know and understand all the evidenceabout which somatic psychological and social factorshave been demonstrated to modulate pain and thenature of their modulation the clinician can considereach factor in terms of what effect it might have onthe implicit perception of threat This conceptualmodel seeks to synthesize that wide body of evidenceinto a principle

Another implication that is worthy of special men-tion is that patients should be helped to base their rea-soning about their condition and their pain onsimilar information This is important because teach-ing patients about modern pain biology leads toaltered beliefs and attitudes about pain40 andincreased pain thresholds during relevant tasks41

Moreover when education about pain biology isincorporated into physiotherapy management ofpatients with chronic pain pain and disability arereduced6869 A key objective of such education is toencourage patients to apply the same principle as thatadvocated for clinicians summarised here as lsquowhateffect might this (factor) have on the implicit percep-tion of threatrsquo or in patient-appropriate languagelsquohow does this affect the answer to the question howdangerous is this reallyrsquo8

USING THIS CONCEPTUALISATION TOUNDERSTAND CRPS AND GUIDE NEW

OPTIONS FOR MANAGEMENT

Complex regional pain syndrome (CRPS) is a debili-tating condition that can occur after minor traumaand sometimes without peripheral trauma for exam-ple post-stroke70 Much is known about the patho-physiology of CRPS including facilitated neurogenic

172 MOSELEY

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

inflammation7172 and tissue hypoxia73 at the injurysite7475 autonomic76 immune77ndash79 motor8081 tactile82ndash85

and proprioceptive86 dysfunction (Fig 2)The syndromic pattern of signs and symptoms

includes pain hyperalgesia allodynia dystoniaswelling abnormal blood flow abnormal sweatinghair and nail growth The sensitivity to provocationcan be remarkable for example elicitation of painswelling and (anecdotally) blood flow changes inresponse to imagined movements87 or when thepatient receives visual input that the limb is beingtouched even though it is not in fact being touched(lsquodysynchiriarsquo)88 The wide-spread and multisystemicnature of the pathophysiology of CRPS implies that

although CRPS is usually initiated by peripheralinsult it is a disorder of the central nervous system75

When one tries to make sense of such a multisystemicand exaggerated response to minor injury the conceptual-isation that pain is a conscious correlate of the implicitperception of the threat to body tissue can be useful Thatpain is just one output by which the brain might try toprotect the tissues ndash one aspect of a homeostaticresponse89 ndash lends itself to CRPS because the otherresponses are so patent That pain is a correlate of implic-itly perceived threat to body tissue rather than the state ofthe tissues or the actual threat to the tissues is particu-larly relevant to CRPS in the absence of any tissue orneural injury for example as a stress response90

RECONCEPTUALISING PAIN ACCORDING TO MODERN PAIN SCIENCE 173

Fig 2 Schematic overview of pathophysiology of complex regional pain syndrome Adapted from Janig and Baron75

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

Each of the pathological findings that have beendocumented in patients with CRPS might be consid-ered a protective response whether it be an immunemotor sensory vascular autonomic or consciousresponse consistent with attempts to protect the partin question by utilising immune motor sensory vas-

cular and autonomic systems as well as conscious-ness Reducing the threshold for activation of theseprotective responses would seem a particularly effec-tive way to protect the body part in question forexample making it so sensitive that even looking at itbeing touched activates a protective response88

174 MOSELEY

Fig 3 Response to different components of motor imagery Pain (A) and functional capacity (B) Recog = laterality recognitionwhereby patients make leftright judgements of pictured hands Imag = imagined movements Mirror = mirror movementsThree groups are shown group 1 undertook the motor imagery program group 2 performed imagined movements first thenlaterality recognition and then mirror movements group 3 performed laterality recognition followed by mirror movements andthen back to laterality recognition Note that group 1 had a largest reduction in pain Note also variable responses to imaginedand mirror movements depending on the order of components From Moseley1 and Moseley93 with permission

Recogniti

on

Neuropathic

pain scale

Task-specific

NRS

Time (weeks)

Time (weeks)

Imag

ined

Mirr

or

A

B

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

The challenge for those trying to understand CRPSaccording to this paradigm is to identify why theimplicit perception of threat to body tissues is soexaggerated in some patients and in some situationsbut not in others Fundamental to the paradigm isthat anything that modulates implicitly perceivedthreat should be relevant That means that psychoso-cial factors including anxiety depression attitudesand beliefs social context or work status may all playan important role Although patients with CRPS donot demonstrate a lsquotypicalrsquo psychosocial profile psy-chosocial contributors are probably relevant in themajority of cases Finally there is initial evidence fora genetic contribution to CRPS91 but more data arerequired to clarify that possibility

Clinical response to CRPS according to this paradigm

If CRPS is an exaggerated protective response then itseems sensible to devise treatment that aims first tofind a baseline that is sufficiently conservative to notelicit the unwanted protective responses (to lsquoget underthe radarrsquo) and second to expose the limb graduallyto threat while continuing to avoid elicitation of theunwanted responses This approach underpins gradedmotor imagery for CRPS9293 whereby patients begintraining by making leftright judgements of picturedlimbs It is known that this task activates cortical net-works that involve representation of the limb andpreparation for movement94 but this task does notactivate primary sensory and motor cortices95

Graded motor imagery progresses from leftright lat-erality judgements to imagined movements which doactivate primary sensory and motor cortices9596 andthen to mirror movements The order of these compo-nents seems to be important in the effect on pain anddisability in patients with chronic CRPS (Fig 3)1 Inpatients with acute (or anecdotally less severe) CRPSit may be sufficient to begin training (conceptualisedhere as exposure to threat) with mirror movements97

One of the key issues outlined earlier is that the ner-vous system changes when nociception and pain persistThere is a large amount of evidence that the cortical rep-resentation of the affected limb undergoes substantialchanges in patients with CRPS4774758398ndash100 and thesechanges have been implicated in the maintenance ofpathological pain syndromes (although see Moseley53

for a word of caution)46 If distorted cortical representa-tion contributes to CRPS then it would seem sensible toattempt to normalise cortical representation of the limbThis has been done in patients with phantom limbpain101 which is associated with changes in primary sen-sory cortex that are probably similar to those observedin CRPS (see Acerra et al102 for a review of common

findings in phantom limb pain stroke and CRPS) Inthat study with amputees sensory discrimination train-ing evoked normalisation of cortical representationimprovement in tactile acuity on the stump and reduc-tionelimination of phantom limb pain101 Increase intactile acuity normalisation of cortical representationand reduction in pain were positively related

Finally if CRPS reflects an exaggerated implicitperception of threat to body tissue then it wouldseem sensible to attempt to reduce the perception ofthreat One approach that has been studied exten-sively in other populations is the explanation to thepatient of the underlying biology of their painPreliminary data from patients with CRPS appearpromising103 but clinical trials are required

CONCLUSIONS

Extensive experimental data corroborate anecdotalevidence that pain does not provide a measure of thestate of the tissues and that pain is modulated bymany factors from across somatic psychological andsocial domains It is now known that as nociceptionand pain persist the neuronal mechanisms involvedin both become more sensitive which means that therelationship between pain and the state of the tissuesbecomes weaker and less predictable One paradigmwhich considers the current thought in pain biologyconceptualises pain as the conscious correlate of theimplicit perception of threat to body tissue This con-ceptualisation can be applied clinically to identify fac-tors from across somatic psychological and socialdomains that may affect the perceived threat to tissuedamage Further it suggests approaches to treatmentthat target those factors Evidence from clinical trialssuggests that clinical strategies based on this concep-tualisation can be effective in patients with disablingcomplex and chronic pain

ACKNOWLEDGEMENTS

The author is supported by the Nuffield DominionsTrust Oxford University UK The conceptual approachfor CRPS draws heavily on unpublished work of MichaelThacker Kingrsquos College London London UK

REFERENCES

1 Moseley GL Is successful rehabilitation of complex regionalpain syndrome due to sustained attention to the affected limbA randomised clinical trial Pain 200511454ndash61

2 Melzack R Wall PD Pain mechanisms a new theory Science1965150971ndash9

RECONCEPTUALISING PAIN ACCORDING TO MODERN PAIN SCIENCE 175

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

and Moerman42 for exhaustive coverage of the role ofmeaning in health and medicine)

To review the very large amount of literature onsomatic psychological and social influences on pain isbeyond the scope of this paper However it is appropriateand clinically meaningful to reiterate the theme thatemerges from that literature that the influences are vari-able and seem to depend on the evaluative context of thenoxious input

The relationship between pain and the state of thetissues becomes weaker as pain persists

The nervous system is dynamic This means that the func-tional properties of individual neurones and of synergiesof neurones change in response to activity To review allthe changes that have been identified is beyond the scopeof this paper and the expertise of this author Howeverthe nature of the changes can be summarised thus thatthe neurones that transmit nociceptive input to the brainbecome sensitised as nociception persists and that thenetworks of neurons within the brain that evoke painbecome sensitised as pain persists The molecular and sys-tems biology of these changes have been discussed at sev-eral levels84344 The clinical manifestations of thesechanges are hyperalgesia (formerly painful stimuli

become more painful) and allodynia (formerly non-painful stimuli become painful) These terms are usedwidely most often in reference to tactile stimuli but alsoin reference to movement and to thermal stimuli

One aspect of the changes that occur when pain persistsis that the proprioceptive representation of the painfulbody part in primary sensory cortex changes45ndash47 Thismay have implications for motor control because theserepresentations are the maps that the brain uses to planand execute movement48 If the map of a body partbecomes inaccurate then motor control may be compro-mised ndash it is known that experimental disruption of corti-cal proprioceptive maps disrupts motor planning49 Thenotion of distorted proprioceptive representation hasbeen discussed with regard to its impact on motor con-trol5051 and more recently in a theoretical way withregard to pain52 Although exceptions exist53 there ismounting evidence that changes in cortical representationoccur in association with chronic pain and it is feasiblethat these changes may become part of the problem46

Conceptualising pain as a conscious correlate of theimplicit perception that tissue is in danger

The biology of pain is complex One response to thiscomplexity is to develop clinically viable conceptual

RECONCEPTUALISING PAIN ACCORDING TO MODERN PAIN SCIENCE 171

Fig 1 Many inputs affect the implicit perception of threat to body tissues labelled here as lsquoHow dangerous is this reallyrsquoThose inputs have wider meaning effects which in turn seems to affect anxiety attention and expectation The implicitperception of threat to body tissues determines pain and in turn influences other inputs

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

paradigms that incorporate what is now known aboutthat complexity One such paradigm that is gaining sup-port is the neuromatrix theory (see Melzack55 for a con-textual review)54 which conceptualises pain as oneoutput of the central nervous system that occurs whenthe organism perceives tissue to be under threat Thereare two important components of this conceptualisa-tion First there are other central nervous system out-puts that occur when tissue is perceived to be underthreat and second that it is the implicit perception ofthreat that determines the outputs not the state of thetissues nor the actual threat to the tissues (Fig 1)

When tissue is under threat a range of local and seg-mental responses occur For example inflammatorymediators are released the body part is usually with-drawn via short and long latency reflex loops there arerapid changes in blood flow and in the excitability ofperipheral nociceptors (so-called peripheral sensitisa-tion)56 The nociceptive system transforms this threatinto electrical activity in peripheral neurones If thismessage of threat is then transmitted by spinal neuronesto higher centres the responses become more complexFor example immune mediators are released into theblood stream57 voluntary and postural muscle activityare altered58 and conscious knowledge of the threat (iepain) may emerge Within this context pain will notemerge until the nociceptive input to the brain has beenevaluated albeit at an unconscious level (see Moseley59

and Gifford et al60 for further discussion)The second important component of the neuromatrix

theory is that pain depends on the perceived degree ofthreat This means that pain can be conceptualised as theconscious correlate of the implicit perception of threat tobody tissues859 That psychosocial factors are very impor-tant in most chronic pain states is well established61ndash65

This paper argues that the mass of data regarding psy-chosocial factors can be gathered within the proposedconceptualisation that pain is one output of the centralnervous system that occurs when the organism perceivestissue to be under threat The conceptualisation has limi-tations and strengths One limitation is that it does notattempt to describe the biology of implicit evaluation ofthreat nor of how this might emerge into consciousnessIn this sense it adds little to theories first proposeddecades ago (see for example Hebb66) However astrength of this conceptualisation is that it can easily beintegrated into a clinical context where making sense ofthe influence of factors from across somatic psychologi-cal and social domains is valuable

Implications for clinical practice

That pain does not reflect the state of the tissues butrather is a conscious driver of behaviour aimed at

protecting those tissues has implications for clinicalpractice One implication is that to base clinical rea-soning on what is currently known about the biologyof pain requires that the skills and knowledge of theclinician are broader than those related to anatomyand biomechanics That is the clinician must have asound knowledge of diagnostic tools tissue dynamicshealing and remodelling peripheral and central sensi-tisation and psychological and social factors thatmight affect the implicit perception of threat to bodytissues This information is readily available and thereis evidence that clinicians can understand modernconcepts with relatively limited training67 That saidit may be unrealistic to expect clinicians to keep up-to-date with progress in knowledge across these areasThis points to a strength of the conceptualisation ofpain as the conscious correlate of the implicit percep-tion threat to body tissues because the clinician canuse the conceptual model to guide treatment That israther than know and understand all the evidenceabout which somatic psychological and social factorshave been demonstrated to modulate pain and thenature of their modulation the clinician can considereach factor in terms of what effect it might have onthe implicit perception of threat This conceptualmodel seeks to synthesize that wide body of evidenceinto a principle

Another implication that is worthy of special men-tion is that patients should be helped to base their rea-soning about their condition and their pain onsimilar information This is important because teach-ing patients about modern pain biology leads toaltered beliefs and attitudes about pain40 andincreased pain thresholds during relevant tasks41

Moreover when education about pain biology isincorporated into physiotherapy management ofpatients with chronic pain pain and disability arereduced6869 A key objective of such education is toencourage patients to apply the same principle as thatadvocated for clinicians summarised here as lsquowhateffect might this (factor) have on the implicit percep-tion of threatrsquo or in patient-appropriate languagelsquohow does this affect the answer to the question howdangerous is this reallyrsquo8

USING THIS CONCEPTUALISATION TOUNDERSTAND CRPS AND GUIDE NEW

OPTIONS FOR MANAGEMENT

Complex regional pain syndrome (CRPS) is a debili-tating condition that can occur after minor traumaand sometimes without peripheral trauma for exam-ple post-stroke70 Much is known about the patho-physiology of CRPS including facilitated neurogenic

172 MOSELEY

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

inflammation7172 and tissue hypoxia73 at the injurysite7475 autonomic76 immune77ndash79 motor8081 tactile82ndash85

and proprioceptive86 dysfunction (Fig 2)The syndromic pattern of signs and symptoms

includes pain hyperalgesia allodynia dystoniaswelling abnormal blood flow abnormal sweatinghair and nail growth The sensitivity to provocationcan be remarkable for example elicitation of painswelling and (anecdotally) blood flow changes inresponse to imagined movements87 or when thepatient receives visual input that the limb is beingtouched even though it is not in fact being touched(lsquodysynchiriarsquo)88 The wide-spread and multisystemicnature of the pathophysiology of CRPS implies that

although CRPS is usually initiated by peripheralinsult it is a disorder of the central nervous system75

When one tries to make sense of such a multisystemicand exaggerated response to minor injury the conceptual-isation that pain is a conscious correlate of the implicitperception of the threat to body tissue can be useful Thatpain is just one output by which the brain might try toprotect the tissues ndash one aspect of a homeostaticresponse89 ndash lends itself to CRPS because the otherresponses are so patent That pain is a correlate of implic-itly perceived threat to body tissue rather than the state ofthe tissues or the actual threat to the tissues is particu-larly relevant to CRPS in the absence of any tissue orneural injury for example as a stress response90

RECONCEPTUALISING PAIN ACCORDING TO MODERN PAIN SCIENCE 173

Fig 2 Schematic overview of pathophysiology of complex regional pain syndrome Adapted from Janig and Baron75

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

Each of the pathological findings that have beendocumented in patients with CRPS might be consid-ered a protective response whether it be an immunemotor sensory vascular autonomic or consciousresponse consistent with attempts to protect the partin question by utilising immune motor sensory vas-

cular and autonomic systems as well as conscious-ness Reducing the threshold for activation of theseprotective responses would seem a particularly effec-tive way to protect the body part in question forexample making it so sensitive that even looking at itbeing touched activates a protective response88

174 MOSELEY

Fig 3 Response to different components of motor imagery Pain (A) and functional capacity (B) Recog = laterality recognitionwhereby patients make leftright judgements of pictured hands Imag = imagined movements Mirror = mirror movementsThree groups are shown group 1 undertook the motor imagery program group 2 performed imagined movements first thenlaterality recognition and then mirror movements group 3 performed laterality recognition followed by mirror movements andthen back to laterality recognition Note that group 1 had a largest reduction in pain Note also variable responses to imaginedand mirror movements depending on the order of components From Moseley1 and Moseley93 with permission

Recogniti

on

Neuropathic

pain scale

Task-specific

NRS

Time (weeks)

Time (weeks)

Imag

ined

Mirr

or

A

B

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

The challenge for those trying to understand CRPSaccording to this paradigm is to identify why theimplicit perception of threat to body tissues is soexaggerated in some patients and in some situationsbut not in others Fundamental to the paradigm isthat anything that modulates implicitly perceivedthreat should be relevant That means that psychoso-cial factors including anxiety depression attitudesand beliefs social context or work status may all playan important role Although patients with CRPS donot demonstrate a lsquotypicalrsquo psychosocial profile psy-chosocial contributors are probably relevant in themajority of cases Finally there is initial evidence fora genetic contribution to CRPS91 but more data arerequired to clarify that possibility

Clinical response to CRPS according to this paradigm

If CRPS is an exaggerated protective response then itseems sensible to devise treatment that aims first tofind a baseline that is sufficiently conservative to notelicit the unwanted protective responses (to lsquoget underthe radarrsquo) and second to expose the limb graduallyto threat while continuing to avoid elicitation of theunwanted responses This approach underpins gradedmotor imagery for CRPS9293 whereby patients begintraining by making leftright judgements of picturedlimbs It is known that this task activates cortical net-works that involve representation of the limb andpreparation for movement94 but this task does notactivate primary sensory and motor cortices95

Graded motor imagery progresses from leftright lat-erality judgements to imagined movements which doactivate primary sensory and motor cortices9596 andthen to mirror movements The order of these compo-nents seems to be important in the effect on pain anddisability in patients with chronic CRPS (Fig 3)1 Inpatients with acute (or anecdotally less severe) CRPSit may be sufficient to begin training (conceptualisedhere as exposure to threat) with mirror movements97

One of the key issues outlined earlier is that the ner-vous system changes when nociception and pain persistThere is a large amount of evidence that the cortical rep-resentation of the affected limb undergoes substantialchanges in patients with CRPS4774758398ndash100 and thesechanges have been implicated in the maintenance ofpathological pain syndromes (although see Moseley53

for a word of caution)46 If distorted cortical representa-tion contributes to CRPS then it would seem sensible toattempt to normalise cortical representation of the limbThis has been done in patients with phantom limbpain101 which is associated with changes in primary sen-sory cortex that are probably similar to those observedin CRPS (see Acerra et al102 for a review of common

findings in phantom limb pain stroke and CRPS) Inthat study with amputees sensory discrimination train-ing evoked normalisation of cortical representationimprovement in tactile acuity on the stump and reduc-tionelimination of phantom limb pain101 Increase intactile acuity normalisation of cortical representationand reduction in pain were positively related

Finally if CRPS reflects an exaggerated implicitperception of threat to body tissue then it wouldseem sensible to attempt to reduce the perception ofthreat One approach that has been studied exten-sively in other populations is the explanation to thepatient of the underlying biology of their painPreliminary data from patients with CRPS appearpromising103 but clinical trials are required

CONCLUSIONS

Extensive experimental data corroborate anecdotalevidence that pain does not provide a measure of thestate of the tissues and that pain is modulated bymany factors from across somatic psychological andsocial domains It is now known that as nociceptionand pain persist the neuronal mechanisms involvedin both become more sensitive which means that therelationship between pain and the state of the tissuesbecomes weaker and less predictable One paradigmwhich considers the current thought in pain biologyconceptualises pain as the conscious correlate of theimplicit perception of threat to body tissue This con-ceptualisation can be applied clinically to identify fac-tors from across somatic psychological and socialdomains that may affect the perceived threat to tissuedamage Further it suggests approaches to treatmentthat target those factors Evidence from clinical trialssuggests that clinical strategies based on this concep-tualisation can be effective in patients with disablingcomplex and chronic pain

ACKNOWLEDGEMENTS

The author is supported by the Nuffield DominionsTrust Oxford University UK The conceptual approachfor CRPS draws heavily on unpublished work of MichaelThacker Kingrsquos College London London UK

REFERENCES

1 Moseley GL Is successful rehabilitation of complex regionalpain syndrome due to sustained attention to the affected limbA randomised clinical trial Pain 200511454ndash61

2 Melzack R Wall PD Pain mechanisms a new theory Science1965150971ndash9

RECONCEPTUALISING PAIN ACCORDING TO MODERN PAIN SCIENCE 175

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

paradigms that incorporate what is now known aboutthat complexity One such paradigm that is gaining sup-port is the neuromatrix theory (see Melzack55 for a con-textual review)54 which conceptualises pain as oneoutput of the central nervous system that occurs whenthe organism perceives tissue to be under threat Thereare two important components of this conceptualisa-tion First there are other central nervous system out-puts that occur when tissue is perceived to be underthreat and second that it is the implicit perception ofthreat that determines the outputs not the state of thetissues nor the actual threat to the tissues (Fig 1)

When tissue is under threat a range of local and seg-mental responses occur For example inflammatorymediators are released the body part is usually with-drawn via short and long latency reflex loops there arerapid changes in blood flow and in the excitability ofperipheral nociceptors (so-called peripheral sensitisa-tion)56 The nociceptive system transforms this threatinto electrical activity in peripheral neurones If thismessage of threat is then transmitted by spinal neuronesto higher centres the responses become more complexFor example immune mediators are released into theblood stream57 voluntary and postural muscle activityare altered58 and conscious knowledge of the threat (iepain) may emerge Within this context pain will notemerge until the nociceptive input to the brain has beenevaluated albeit at an unconscious level (see Moseley59

and Gifford et al60 for further discussion)The second important component of the neuromatrix

theory is that pain depends on the perceived degree ofthreat This means that pain can be conceptualised as theconscious correlate of the implicit perception of threat tobody tissues859 That psychosocial factors are very impor-tant in most chronic pain states is well established61ndash65

This paper argues that the mass of data regarding psy-chosocial factors can be gathered within the proposedconceptualisation that pain is one output of the centralnervous system that occurs when the organism perceivestissue to be under threat The conceptualisation has limi-tations and strengths One limitation is that it does notattempt to describe the biology of implicit evaluation ofthreat nor of how this might emerge into consciousnessIn this sense it adds little to theories first proposeddecades ago (see for example Hebb66) However astrength of this conceptualisation is that it can easily beintegrated into a clinical context where making sense ofthe influence of factors from across somatic psychologi-cal and social domains is valuable

Implications for clinical practice

That pain does not reflect the state of the tissues butrather is a conscious driver of behaviour aimed at

protecting those tissues has implications for clinicalpractice One implication is that to base clinical rea-soning on what is currently known about the biologyof pain requires that the skills and knowledge of theclinician are broader than those related to anatomyand biomechanics That is the clinician must have asound knowledge of diagnostic tools tissue dynamicshealing and remodelling peripheral and central sensi-tisation and psychological and social factors thatmight affect the implicit perception of threat to bodytissues This information is readily available and thereis evidence that clinicians can understand modernconcepts with relatively limited training67 That saidit may be unrealistic to expect clinicians to keep up-to-date with progress in knowledge across these areasThis points to a strength of the conceptualisation ofpain as the conscious correlate of the implicit percep-tion threat to body tissues because the clinician canuse the conceptual model to guide treatment That israther than know and understand all the evidenceabout which somatic psychological and social factorshave been demonstrated to modulate pain and thenature of their modulation the clinician can considereach factor in terms of what effect it might have onthe implicit perception of threat This conceptualmodel seeks to synthesize that wide body of evidenceinto a principle

Another implication that is worthy of special men-tion is that patients should be helped to base their rea-soning about their condition and their pain onsimilar information This is important because teach-ing patients about modern pain biology leads toaltered beliefs and attitudes about pain40 andincreased pain thresholds during relevant tasks41

Moreover when education about pain biology isincorporated into physiotherapy management ofpatients with chronic pain pain and disability arereduced6869 A key objective of such education is toencourage patients to apply the same principle as thatadvocated for clinicians summarised here as lsquowhateffect might this (factor) have on the implicit percep-tion of threatrsquo or in patient-appropriate languagelsquohow does this affect the answer to the question howdangerous is this reallyrsquo8

USING THIS CONCEPTUALISATION TOUNDERSTAND CRPS AND GUIDE NEW

OPTIONS FOR MANAGEMENT

Complex regional pain syndrome (CRPS) is a debili-tating condition that can occur after minor traumaand sometimes without peripheral trauma for exam-ple post-stroke70 Much is known about the patho-physiology of CRPS including facilitated neurogenic

172 MOSELEY

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

inflammation7172 and tissue hypoxia73 at the injurysite7475 autonomic76 immune77ndash79 motor8081 tactile82ndash85

and proprioceptive86 dysfunction (Fig 2)The syndromic pattern of signs and symptoms

includes pain hyperalgesia allodynia dystoniaswelling abnormal blood flow abnormal sweatinghair and nail growth The sensitivity to provocationcan be remarkable for example elicitation of painswelling and (anecdotally) blood flow changes inresponse to imagined movements87 or when thepatient receives visual input that the limb is beingtouched even though it is not in fact being touched(lsquodysynchiriarsquo)88 The wide-spread and multisystemicnature of the pathophysiology of CRPS implies that

although CRPS is usually initiated by peripheralinsult it is a disorder of the central nervous system75

When one tries to make sense of such a multisystemicand exaggerated response to minor injury the conceptual-isation that pain is a conscious correlate of the implicitperception of the threat to body tissue can be useful Thatpain is just one output by which the brain might try toprotect the tissues ndash one aspect of a homeostaticresponse89 ndash lends itself to CRPS because the otherresponses are so patent That pain is a correlate of implic-itly perceived threat to body tissue rather than the state ofthe tissues or the actual threat to the tissues is particu-larly relevant to CRPS in the absence of any tissue orneural injury for example as a stress response90

RECONCEPTUALISING PAIN ACCORDING TO MODERN PAIN SCIENCE 173

Fig 2 Schematic overview of pathophysiology of complex regional pain syndrome Adapted from Janig and Baron75

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

Each of the pathological findings that have beendocumented in patients with CRPS might be consid-ered a protective response whether it be an immunemotor sensory vascular autonomic or consciousresponse consistent with attempts to protect the partin question by utilising immune motor sensory vas-

cular and autonomic systems as well as conscious-ness Reducing the threshold for activation of theseprotective responses would seem a particularly effec-tive way to protect the body part in question forexample making it so sensitive that even looking at itbeing touched activates a protective response88

174 MOSELEY

Fig 3 Response to different components of motor imagery Pain (A) and functional capacity (B) Recog = laterality recognitionwhereby patients make leftright judgements of pictured hands Imag = imagined movements Mirror = mirror movementsThree groups are shown group 1 undertook the motor imagery program group 2 performed imagined movements first thenlaterality recognition and then mirror movements group 3 performed laterality recognition followed by mirror movements andthen back to laterality recognition Note that group 1 had a largest reduction in pain Note also variable responses to imaginedand mirror movements depending on the order of components From Moseley1 and Moseley93 with permission

Recogniti

on

Neuropathic

pain scale

Task-specific

NRS

Time (weeks)

Time (weeks)

Imag

ined

Mirr

or

A

B

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

The challenge for those trying to understand CRPSaccording to this paradigm is to identify why theimplicit perception of threat to body tissues is soexaggerated in some patients and in some situationsbut not in others Fundamental to the paradigm isthat anything that modulates implicitly perceivedthreat should be relevant That means that psychoso-cial factors including anxiety depression attitudesand beliefs social context or work status may all playan important role Although patients with CRPS donot demonstrate a lsquotypicalrsquo psychosocial profile psy-chosocial contributors are probably relevant in themajority of cases Finally there is initial evidence fora genetic contribution to CRPS91 but more data arerequired to clarify that possibility

Clinical response to CRPS according to this paradigm

If CRPS is an exaggerated protective response then itseems sensible to devise treatment that aims first tofind a baseline that is sufficiently conservative to notelicit the unwanted protective responses (to lsquoget underthe radarrsquo) and second to expose the limb graduallyto threat while continuing to avoid elicitation of theunwanted responses This approach underpins gradedmotor imagery for CRPS9293 whereby patients begintraining by making leftright judgements of picturedlimbs It is known that this task activates cortical net-works that involve representation of the limb andpreparation for movement94 but this task does notactivate primary sensory and motor cortices95

Graded motor imagery progresses from leftright lat-erality judgements to imagined movements which doactivate primary sensory and motor cortices9596 andthen to mirror movements The order of these compo-nents seems to be important in the effect on pain anddisability in patients with chronic CRPS (Fig 3)1 Inpatients with acute (or anecdotally less severe) CRPSit may be sufficient to begin training (conceptualisedhere as exposure to threat) with mirror movements97

One of the key issues outlined earlier is that the ner-vous system changes when nociception and pain persistThere is a large amount of evidence that the cortical rep-resentation of the affected limb undergoes substantialchanges in patients with CRPS4774758398ndash100 and thesechanges have been implicated in the maintenance ofpathological pain syndromes (although see Moseley53

for a word of caution)46 If distorted cortical representa-tion contributes to CRPS then it would seem sensible toattempt to normalise cortical representation of the limbThis has been done in patients with phantom limbpain101 which is associated with changes in primary sen-sory cortex that are probably similar to those observedin CRPS (see Acerra et al102 for a review of common

findings in phantom limb pain stroke and CRPS) Inthat study with amputees sensory discrimination train-ing evoked normalisation of cortical representationimprovement in tactile acuity on the stump and reduc-tionelimination of phantom limb pain101 Increase intactile acuity normalisation of cortical representationand reduction in pain were positively related

Finally if CRPS reflects an exaggerated implicitperception of threat to body tissue then it wouldseem sensible to attempt to reduce the perception ofthreat One approach that has been studied exten-sively in other populations is the explanation to thepatient of the underlying biology of their painPreliminary data from patients with CRPS appearpromising103 but clinical trials are required

CONCLUSIONS

Extensive experimental data corroborate anecdotalevidence that pain does not provide a measure of thestate of the tissues and that pain is modulated bymany factors from across somatic psychological andsocial domains It is now known that as nociceptionand pain persist the neuronal mechanisms involvedin both become more sensitive which means that therelationship between pain and the state of the tissuesbecomes weaker and less predictable One paradigmwhich considers the current thought in pain biologyconceptualises pain as the conscious correlate of theimplicit perception of threat to body tissue This con-ceptualisation can be applied clinically to identify fac-tors from across somatic psychological and socialdomains that may affect the perceived threat to tissuedamage Further it suggests approaches to treatmentthat target those factors Evidence from clinical trialssuggests that clinical strategies based on this concep-tualisation can be effective in patients with disablingcomplex and chronic pain

ACKNOWLEDGEMENTS

The author is supported by the Nuffield DominionsTrust Oxford University UK The conceptual approachfor CRPS draws heavily on unpublished work of MichaelThacker Kingrsquos College London London UK

REFERENCES

1 Moseley GL Is successful rehabilitation of complex regionalpain syndrome due to sustained attention to the affected limbA randomised clinical trial Pain 200511454ndash61

2 Melzack R Wall PD Pain mechanisms a new theory Science1965150971ndash9

RECONCEPTUALISING PAIN ACCORDING TO MODERN PAIN SCIENCE 175

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

inflammation7172 and tissue hypoxia73 at the injurysite7475 autonomic76 immune77ndash79 motor8081 tactile82ndash85

and proprioceptive86 dysfunction (Fig 2)The syndromic pattern of signs and symptoms

includes pain hyperalgesia allodynia dystoniaswelling abnormal blood flow abnormal sweatinghair and nail growth The sensitivity to provocationcan be remarkable for example elicitation of painswelling and (anecdotally) blood flow changes inresponse to imagined movements87 or when thepatient receives visual input that the limb is beingtouched even though it is not in fact being touched(lsquodysynchiriarsquo)88 The wide-spread and multisystemicnature of the pathophysiology of CRPS implies that

although CRPS is usually initiated by peripheralinsult it is a disorder of the central nervous system75

When one tries to make sense of such a multisystemicand exaggerated response to minor injury the conceptual-isation that pain is a conscious correlate of the implicitperception of the threat to body tissue can be useful Thatpain is just one output by which the brain might try toprotect the tissues ndash one aspect of a homeostaticresponse89 ndash lends itself to CRPS because the otherresponses are so patent That pain is a correlate of implic-itly perceived threat to body tissue rather than the state ofthe tissues or the actual threat to the tissues is particu-larly relevant to CRPS in the absence of any tissue orneural injury for example as a stress response90

RECONCEPTUALISING PAIN ACCORDING TO MODERN PAIN SCIENCE 173

Fig 2 Schematic overview of pathophysiology of complex regional pain syndrome Adapted from Janig and Baron75

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

Each of the pathological findings that have beendocumented in patients with CRPS might be consid-ered a protective response whether it be an immunemotor sensory vascular autonomic or consciousresponse consistent with attempts to protect the partin question by utilising immune motor sensory vas-

cular and autonomic systems as well as conscious-ness Reducing the threshold for activation of theseprotective responses would seem a particularly effec-tive way to protect the body part in question forexample making it so sensitive that even looking at itbeing touched activates a protective response88

174 MOSELEY

Fig 3 Response to different components of motor imagery Pain (A) and functional capacity (B) Recog = laterality recognitionwhereby patients make leftright judgements of pictured hands Imag = imagined movements Mirror = mirror movementsThree groups are shown group 1 undertook the motor imagery program group 2 performed imagined movements first thenlaterality recognition and then mirror movements group 3 performed laterality recognition followed by mirror movements andthen back to laterality recognition Note that group 1 had a largest reduction in pain Note also variable responses to imaginedand mirror movements depending on the order of components From Moseley1 and Moseley93 with permission

Recogniti

on

Neuropathic

pain scale

Task-specific

NRS

Time (weeks)

Time (weeks)

Imag

ined

Mirr

or

A

B

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

The challenge for those trying to understand CRPSaccording to this paradigm is to identify why theimplicit perception of threat to body tissues is soexaggerated in some patients and in some situationsbut not in others Fundamental to the paradigm isthat anything that modulates implicitly perceivedthreat should be relevant That means that psychoso-cial factors including anxiety depression attitudesand beliefs social context or work status may all playan important role Although patients with CRPS donot demonstrate a lsquotypicalrsquo psychosocial profile psy-chosocial contributors are probably relevant in themajority of cases Finally there is initial evidence fora genetic contribution to CRPS91 but more data arerequired to clarify that possibility

Clinical response to CRPS according to this paradigm

If CRPS is an exaggerated protective response then itseems sensible to devise treatment that aims first tofind a baseline that is sufficiently conservative to notelicit the unwanted protective responses (to lsquoget underthe radarrsquo) and second to expose the limb graduallyto threat while continuing to avoid elicitation of theunwanted responses This approach underpins gradedmotor imagery for CRPS9293 whereby patients begintraining by making leftright judgements of picturedlimbs It is known that this task activates cortical net-works that involve representation of the limb andpreparation for movement94 but this task does notactivate primary sensory and motor cortices95

Graded motor imagery progresses from leftright lat-erality judgements to imagined movements which doactivate primary sensory and motor cortices9596 andthen to mirror movements The order of these compo-nents seems to be important in the effect on pain anddisability in patients with chronic CRPS (Fig 3)1 Inpatients with acute (or anecdotally less severe) CRPSit may be sufficient to begin training (conceptualisedhere as exposure to threat) with mirror movements97

One of the key issues outlined earlier is that the ner-vous system changes when nociception and pain persistThere is a large amount of evidence that the cortical rep-resentation of the affected limb undergoes substantialchanges in patients with CRPS4774758398ndash100 and thesechanges have been implicated in the maintenance ofpathological pain syndromes (although see Moseley53

for a word of caution)46 If distorted cortical representa-tion contributes to CRPS then it would seem sensible toattempt to normalise cortical representation of the limbThis has been done in patients with phantom limbpain101 which is associated with changes in primary sen-sory cortex that are probably similar to those observedin CRPS (see Acerra et al102 for a review of common

findings in phantom limb pain stroke and CRPS) Inthat study with amputees sensory discrimination train-ing evoked normalisation of cortical representationimprovement in tactile acuity on the stump and reduc-tionelimination of phantom limb pain101 Increase intactile acuity normalisation of cortical representationand reduction in pain were positively related

Finally if CRPS reflects an exaggerated implicitperception of threat to body tissue then it wouldseem sensible to attempt to reduce the perception ofthreat One approach that has been studied exten-sively in other populations is the explanation to thepatient of the underlying biology of their painPreliminary data from patients with CRPS appearpromising103 but clinical trials are required

CONCLUSIONS

Extensive experimental data corroborate anecdotalevidence that pain does not provide a measure of thestate of the tissues and that pain is modulated bymany factors from across somatic psychological andsocial domains It is now known that as nociceptionand pain persist the neuronal mechanisms involvedin both become more sensitive which means that therelationship between pain and the state of the tissuesbecomes weaker and less predictable One paradigmwhich considers the current thought in pain biologyconceptualises pain as the conscious correlate of theimplicit perception of threat to body tissue This con-ceptualisation can be applied clinically to identify fac-tors from across somatic psychological and socialdomains that may affect the perceived threat to tissuedamage Further it suggests approaches to treatmentthat target those factors Evidence from clinical trialssuggests that clinical strategies based on this concep-tualisation can be effective in patients with disablingcomplex and chronic pain

ACKNOWLEDGEMENTS

The author is supported by the Nuffield DominionsTrust Oxford University UK The conceptual approachfor CRPS draws heavily on unpublished work of MichaelThacker Kingrsquos College London London UK

REFERENCES

1 Moseley GL Is successful rehabilitation of complex regionalpain syndrome due to sustained attention to the affected limbA randomised clinical trial Pain 200511454ndash61

2 Melzack R Wall PD Pain mechanisms a new theory Science1965150971ndash9

RECONCEPTUALISING PAIN ACCORDING TO MODERN PAIN SCIENCE 175

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

Each of the pathological findings that have beendocumented in patients with CRPS might be consid-ered a protective response whether it be an immunemotor sensory vascular autonomic or consciousresponse consistent with attempts to protect the partin question by utilising immune motor sensory vas-

cular and autonomic systems as well as conscious-ness Reducing the threshold for activation of theseprotective responses would seem a particularly effec-tive way to protect the body part in question forexample making it so sensitive that even looking at itbeing touched activates a protective response88

174 MOSELEY

Fig 3 Response to different components of motor imagery Pain (A) and functional capacity (B) Recog = laterality recognitionwhereby patients make leftright judgements of pictured hands Imag = imagined movements Mirror = mirror movementsThree groups are shown group 1 undertook the motor imagery program group 2 performed imagined movements first thenlaterality recognition and then mirror movements group 3 performed laterality recognition followed by mirror movements andthen back to laterality recognition Note that group 1 had a largest reduction in pain Note also variable responses to imaginedand mirror movements depending on the order of components From Moseley1 and Moseley93 with permission

Recogniti

on

Neuropathic

pain scale

Task-specific

NRS

Time (weeks)

Time (weeks)

Imag

ined

Mirr

or

A

B

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

The challenge for those trying to understand CRPSaccording to this paradigm is to identify why theimplicit perception of threat to body tissues is soexaggerated in some patients and in some situationsbut not in others Fundamental to the paradigm isthat anything that modulates implicitly perceivedthreat should be relevant That means that psychoso-cial factors including anxiety depression attitudesand beliefs social context or work status may all playan important role Although patients with CRPS donot demonstrate a lsquotypicalrsquo psychosocial profile psy-chosocial contributors are probably relevant in themajority of cases Finally there is initial evidence fora genetic contribution to CRPS91 but more data arerequired to clarify that possibility

Clinical response to CRPS according to this paradigm

If CRPS is an exaggerated protective response then itseems sensible to devise treatment that aims first tofind a baseline that is sufficiently conservative to notelicit the unwanted protective responses (to lsquoget underthe radarrsquo) and second to expose the limb graduallyto threat while continuing to avoid elicitation of theunwanted responses This approach underpins gradedmotor imagery for CRPS9293 whereby patients begintraining by making leftright judgements of picturedlimbs It is known that this task activates cortical net-works that involve representation of the limb andpreparation for movement94 but this task does notactivate primary sensory and motor cortices95

Graded motor imagery progresses from leftright lat-erality judgements to imagined movements which doactivate primary sensory and motor cortices9596 andthen to mirror movements The order of these compo-nents seems to be important in the effect on pain anddisability in patients with chronic CRPS (Fig 3)1 Inpatients with acute (or anecdotally less severe) CRPSit may be sufficient to begin training (conceptualisedhere as exposure to threat) with mirror movements97

One of the key issues outlined earlier is that the ner-vous system changes when nociception and pain persistThere is a large amount of evidence that the cortical rep-resentation of the affected limb undergoes substantialchanges in patients with CRPS4774758398ndash100 and thesechanges have been implicated in the maintenance ofpathological pain syndromes (although see Moseley53

for a word of caution)46 If distorted cortical representa-tion contributes to CRPS then it would seem sensible toattempt to normalise cortical representation of the limbThis has been done in patients with phantom limbpain101 which is associated with changes in primary sen-sory cortex that are probably similar to those observedin CRPS (see Acerra et al102 for a review of common

findings in phantom limb pain stroke and CRPS) Inthat study with amputees sensory discrimination train-ing evoked normalisation of cortical representationimprovement in tactile acuity on the stump and reduc-tionelimination of phantom limb pain101 Increase intactile acuity normalisation of cortical representationand reduction in pain were positively related

Finally if CRPS reflects an exaggerated implicitperception of threat to body tissue then it wouldseem sensible to attempt to reduce the perception ofthreat One approach that has been studied exten-sively in other populations is the explanation to thepatient of the underlying biology of their painPreliminary data from patients with CRPS appearpromising103 but clinical trials are required

CONCLUSIONS

Extensive experimental data corroborate anecdotalevidence that pain does not provide a measure of thestate of the tissues and that pain is modulated bymany factors from across somatic psychological andsocial domains It is now known that as nociceptionand pain persist the neuronal mechanisms involvedin both become more sensitive which means that therelationship between pain and the state of the tissuesbecomes weaker and less predictable One paradigmwhich considers the current thought in pain biologyconceptualises pain as the conscious correlate of theimplicit perception of threat to body tissue This con-ceptualisation can be applied clinically to identify fac-tors from across somatic psychological and socialdomains that may affect the perceived threat to tissuedamage Further it suggests approaches to treatmentthat target those factors Evidence from clinical trialssuggests that clinical strategies based on this concep-tualisation can be effective in patients with disablingcomplex and chronic pain

ACKNOWLEDGEMENTS

The author is supported by the Nuffield DominionsTrust Oxford University UK The conceptual approachfor CRPS draws heavily on unpublished work of MichaelThacker Kingrsquos College London London UK

REFERENCES

1 Moseley GL Is successful rehabilitation of complex regionalpain syndrome due to sustained attention to the affected limbA randomised clinical trial Pain 200511454ndash61

2 Melzack R Wall PD Pain mechanisms a new theory Science1965150971ndash9

RECONCEPTUALISING PAIN ACCORDING TO MODERN PAIN SCIENCE 175

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

The challenge for those trying to understand CRPSaccording to this paradigm is to identify why theimplicit perception of threat to body tissues is soexaggerated in some patients and in some situationsbut not in others Fundamental to the paradigm isthat anything that modulates implicitly perceivedthreat should be relevant That means that psychoso-cial factors including anxiety depression attitudesand beliefs social context or work status may all playan important role Although patients with CRPS donot demonstrate a lsquotypicalrsquo psychosocial profile psy-chosocial contributors are probably relevant in themajority of cases Finally there is initial evidence fora genetic contribution to CRPS91 but more data arerequired to clarify that possibility

Clinical response to CRPS according to this paradigm

If CRPS is an exaggerated protective response then itseems sensible to devise treatment that aims first tofind a baseline that is sufficiently conservative to notelicit the unwanted protective responses (to lsquoget underthe radarrsquo) and second to expose the limb graduallyto threat while continuing to avoid elicitation of theunwanted responses This approach underpins gradedmotor imagery for CRPS9293 whereby patients begintraining by making leftright judgements of picturedlimbs It is known that this task activates cortical net-works that involve representation of the limb andpreparation for movement94 but this task does notactivate primary sensory and motor cortices95

Graded motor imagery progresses from leftright lat-erality judgements to imagined movements which doactivate primary sensory and motor cortices9596 andthen to mirror movements The order of these compo-nents seems to be important in the effect on pain anddisability in patients with chronic CRPS (Fig 3)1 Inpatients with acute (or anecdotally less severe) CRPSit may be sufficient to begin training (conceptualisedhere as exposure to threat) with mirror movements97

One of the key issues outlined earlier is that the ner-vous system changes when nociception and pain persistThere is a large amount of evidence that the cortical rep-resentation of the affected limb undergoes substantialchanges in patients with CRPS4774758398ndash100 and thesechanges have been implicated in the maintenance ofpathological pain syndromes (although see Moseley53

for a word of caution)46 If distorted cortical representa-tion contributes to CRPS then it would seem sensible toattempt to normalise cortical representation of the limbThis has been done in patients with phantom limbpain101 which is associated with changes in primary sen-sory cortex that are probably similar to those observedin CRPS (see Acerra et al102 for a review of common

findings in phantom limb pain stroke and CRPS) Inthat study with amputees sensory discrimination train-ing evoked normalisation of cortical representationimprovement in tactile acuity on the stump and reduc-tionelimination of phantom limb pain101 Increase intactile acuity normalisation of cortical representationand reduction in pain were positively related

Finally if CRPS reflects an exaggerated implicitperception of threat to body tissue then it wouldseem sensible to attempt to reduce the perception ofthreat One approach that has been studied exten-sively in other populations is the explanation to thepatient of the underlying biology of their painPreliminary data from patients with CRPS appearpromising103 but clinical trials are required

CONCLUSIONS

Extensive experimental data corroborate anecdotalevidence that pain does not provide a measure of thestate of the tissues and that pain is modulated bymany factors from across somatic psychological andsocial domains It is now known that as nociceptionand pain persist the neuronal mechanisms involvedin both become more sensitive which means that therelationship between pain and the state of the tissuesbecomes weaker and less predictable One paradigmwhich considers the current thought in pain biologyconceptualises pain as the conscious correlate of theimplicit perception of threat to body tissue This con-ceptualisation can be applied clinically to identify fac-tors from across somatic psychological and socialdomains that may affect the perceived threat to tissuedamage Further it suggests approaches to treatmentthat target those factors Evidence from clinical trialssuggests that clinical strategies based on this concep-tualisation can be effective in patients with disablingcomplex and chronic pain

ACKNOWLEDGEMENTS

The author is supported by the Nuffield DominionsTrust Oxford University UK The conceptual approachfor CRPS draws heavily on unpublished work of MichaelThacker Kingrsquos College London London UK

REFERENCES

1 Moseley GL Is successful rehabilitation of complex regionalpain syndrome due to sustained attention to the affected limbA randomised clinical trial Pain 200511454ndash61

2 Melzack R Wall PD Pain mechanisms a new theory Science1965150971ndash9

RECONCEPTUALISING PAIN ACCORDING TO MODERN PAIN SCIENCE 175

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

3 Wall PD McMahon SB The relationship of perceived pain toafferent nerve impulses Trends Neurosci 19869254ndash5

4 Vierck J Charles J Animal models of pain In McMahon SBKoltzenburg M (eds) Textbook of Pain 5th edn LondonElsevier 2006 175ndash85

5 Handwerker HO Kobal G Psychophysiology ofexperimentally induced pain Physiol Rev 199373639ndash71

6 Meyer R Ringkamp M Campbell JN Raja SN Peripheralmechanisms of cutaneous nociception In McMahon SBKoltzenburg M (eds) Textbook of Pain 5th edn LondonElsevier 2006 3ndash35

7 Craig A How do you feel Interoception the sense of thephysiological condition of the body Nat Rev Neurosci20023655ndash66

8 Butler D Moseley GL Explain pain Adelaide NOI GroupPublishing 2003

9 Fields H Basbaum A Heinricher M CNS mechanisms ofpain modulation In McMahon SB Koltzenburg M (eds)Textbook of Pain 5th edn London Elsevier 2006 125ndash43

10 Asmundson GJ Kuperos JL Norton GR Do patients withchronic pain selectively attend to pain-related informationPreliminary evidence for the mediating role of fear Pain19977227ndash32

11 Crombez G Eccleston C Baeyens F Eelen P The disruptivenature of pain an experimental investigation Behav Res Ther199634911ndash8

12 Crombez G Eccleston C Baeyens F Eelen P Habituation andthe interference of pain with task performance Pain199770149ndash54

13 Crombez G Eccleston C Baeyens F Eelen P Attentionaldisruption is enhanced by the threat of pain Behav Res Ther199836195ndash204

14 Duckworth MP Iezzi A Adams HE Hale D Informationprocessing in chronic pain disorder a preliminary analysis JPsychopathol Behav Assess 199719239ndash55

15 Eccleston C Chronic pain and attention a cognitive approachBr J Clin Psychol 199433535ndash47

16 Eccleston C Crombez G Aldrich S Stannard C Attentionand somatic awareness in chronic pain Pain 199772209ndash15

17 Matthews KA Schier MF Brunson BI Carducci B Attentionunpredictability and reports of physical symptoms eliminatingthe benefits of predictability J Personal Soc Psychol198038525ndash37

18 McCracken LM lsquoAttentionrsquo to pain in persons with chronicpain a behavioral approach Behav Ther 199728271ndash84

19 Peters ML Vlaeyen JW van Drunen C Do fibromyalgiapatients display hypervigilance for innocuous somatosensorystimuli Application of a body scanning reaction timeparadigm Pain 200086283ndash92

20 Crombez G Eccleston C Baeyens F van Houdenhove B vanden Broeck A Attention to chronic pain is dependent uponpain-related fear J Psychosom Res 199947403ndash10

21 Eccleston C Crombez G Pain demands attention a cognitive-affective model of the interruptive function of pain PsycholBull 1999125356ndash66

22 Moseley GL Arntz A The context of a noxious stimulusaffects the pain it evokes Pain 2007 In press

23 Klages U Kianifard S Ulusoy O Wehrbein H Anxietysensitivity as predictor of pain in patients undergoingrestorative dental procedures Community Dent Oral Epidemiol200634139ndash45

24 Schupp CJ Berbaum K Berbaum M Lang EV Pain andanxiety during interventional radiologic procedures effect ofpatientsrsquo state anxiety at baseline and modulation bynonpharmacologic analgesia adjuncts J Vasc Intervent Radiol2005161585ndash92

25 Gore M Brandenburg NA Dukes E Hoffman DL Tai KSStacey B Pain severity in diabetic peripheral neuropathy isassociated with patient functioning symptom levels of anxietyand depression and sleep J Pain Symptom Manage200530374ndash85

26 Pud D Amit A Anxiety as a predictor of pain magnitudefollowing termination of first-trimester pregnancy Pain Med20056143ndash8

27 Tang J Gibson SJ A psychophysical evaluation of therelationship between trait anxiety pain perception andinduced state anxiety J Pain 20056612ndash9

28 Arntz A Dreessen L De Jong P The influence of anxiety onpain Attentional and attributional mediators Pain199456307ndash14

29 Arntz A Vaneck M Heijmans M Predictions of dental pain ndashthe fear of any expected evil is worse than the evil itself BehavRes Ther 19902829ndash41

30 Ploghaus A Becerra L Borras C Borsook D Neural circuitryunderlying pain modulation expectation hypnosis placeboTrends Cognitive Sci 20037197ndash200

31 Sawamoto N Honda M Okada T Hanakawa T Kanda MFukuyama H et al Expectation of pain enhances responses tononpainful somatosensory stimulation in the anteriorcingulate cortex and parietal operculumposterior insula anevent-related functional magnetic resonance imaging study JNeurosci 2000207438ndash45

32 Ploghaus A Tracey I Gati JS Clare S Menon RS MatthewsPM et al Dissociating pain from its anticipation in the humanbrain Science 19992841979ndash81

33 Chua P Krams M Toni I Passingham R Dolan R Afunctional anatomy of anticipatory anxiety Neuroimage19999563ndash71

34 Fields HL Pain modulation expectation opioid analgesiaand virtual pain Biol Basis Mind Body Interact 2000245ndash53

35 Keltner JR Furst A Fan C Redfern R Inglis B Fields HLIsolating the modulatory effect of expectation on paintransmission a functional magnetic resonance imaging studyJ Neurosci 2006264437ndash43

36 Wager TD Expectations and anxiety as mediators of placeboeffects in pain Pain 2005115225ndash6

37 Benedetti F Pollo A Lopiano L Lanotte M Vighetti SRainero I Conscious expectation and unconsciousconditioning in analgesic motor and hormonalplacebonocebo responses J Neurosci 2003234315ndash23

38 Pollo A Amanzio M Arslanian A Casadio C Maggi GBenedetti F Response expectancies in placebo analgesia andtheir clinical relevance Pain 20019377ndash84

39 Sullivan MJL Thorn B Haythornthwaite JA Keefe F MartinM Bradley LA et al Theoretical perspectives on the relationbetween catastrophizing and pain Clin J Pain 20011752ndash64

40 Moseley GL Nicholas MK Hodges PW A randomizedcontrolled trial of intensive neurophysiology education inchronic low back pain Clin J Pain 200420324ndash30

41 Moseley GL Evidence for a direct relationship betweencognitive and physical change during an educationintervention in people with chronic low back pain Eur J Pain2004839ndash45

42 Moerman D Meaning medicine and the lsquoplacebo effectrsquoCambridge UK Cambridge University Press 2002

43 McMahon SB Koltzenburg M Wall and Melzackrsquos Textbookof Pain London Elsevier 2006

44 Butler D The sensitive nervous system Adelaide NOIPublications 2000

45 Flor H Braun C Elbert T Birbaumer N Extensivereorganization of primary somatosensory cortex in chronicback pain patients Neurosci Lett 19972245ndash8

176 MOSELEY

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

46 Flor H Nikolajsen L Jensen TS Phantom limb pain a case ofmaladaptive CNS plasticity Nat Rev Neurosci 20067873ndash81

47 Maihofner C Handwerker HO Neundorfer B Birklein FPatterns of cortical reorganization in complex regional painsyndrome Neurology 2003611707ndash15

48 Buonomano D Merzenich M Cortical plasticity fromsynapses to maps Annu Rev Neurosci 199821149ndash86

49 McCormick K Zalucki N Hudson M Moseley GL Doesfaulty proprioceptive input disrupt motor imagery Aus JPhysiother 2007 In press

50 Byl NN McKenzie A Nagarajan SS Differences insomatosensory hand organization in a healthy flutist and aflutist with focal hand dystonia a case report J Hand Ther200013302ndash9

51 Byl NN Merzenich MM Cheung S Bedenbaugh PNagarajan SS Jenkins WM A primate model for studyingfocal dystonia and repetitive strain injury effects on theprimary somatosensory cortex Phys Ther 199777269ndash84

52 Harris AJ Cortical origin of pathological pain Lancet19993541464ndash6

53 Moseley GL Making sense of S1 mania ndash are things really thatsimple In Gifford L (ed) Topical Issues in Pain volume 5Falmouth CNS Press 2006 321ndash40

54 Melzack R Phantom limbs and the concept of a neuromatrixTrends Neurosci 19901388ndash92

55 Melzack R Gate control theory On the evolution of painconcepts Pain Forum 19965128ndash38

56 Bevan S Nociceptive peripheral neurons cellular propertiesIn Wall P Melzack R (eds) The textbook of pain 4th ednEdinburgh Churchill Livingstone 1999 85ndash103

57 Watkins L Maier S The pain of being sick implications ofimmune-to-brain communication for understanding painAnnu Rev Psychol 20005129ndash57

58 Hodges PW Moseley GL Pain and motor control of thelumbopelvic region effect and possible mechanisms JElectromyogr Kinesiol 200313361ndash70

59 Moseley GL A pain neuromatrix approach to patients withchronic pain Man Ther 20038130ndash40

60 Gifford L Thacker M Jones M Physiotherapy and pain InMcMahon SB Koltzenburg M (eds) Textbook of PainLondon Elsevier 2006 603ndash18

61 Turner JA Jensen MP Romano JM Do beliefs coping andcatastrophizing independently predict functioning in patientswith chronic pain Pain 200085115ndash25

62 Stroud MW Thorn BE Jensen MP Boothby JL The relationbetween pain beliefs negative thoughts and psychosocialfunctioning in chronic pain patients Pain 200084347ndash52

63 Turk DC Flor H Chronic pain a biobehavioral perspectiveIn Gatchel R Turk DC (eds) Psychosocial factors in painCritical perspectives New York Guilford 1999 18ndash34

64 Kendall NAS Linton SJ Main C Psychosocial yellow flagsfor acute low back pain lsquoyellow flagsrsquo as an analogue to lsquoredflagsrsquo Eur J Pain 1998287ndash9

65 Linton SJ A review of psychological risk factors in back andneck pain Spine 2000251148ndash56

66 Hebb D The organization of behavior New York Wiley 194967 Moseley GL Unravelling the barriers to reconceptualisation

of the problem in chronic pain the actual and perceived abilityof patients and health professionals to understand theneurophysiology J Pain 20034184ndash9

68 Moseley GL Combined physiotherapy and education iseffective for chronic low back pain A randomised controlledtrial Aus J Physiother 200248297ndash302

69 Moseley GL Joining forces ndash combining cognition-targetedmotor control training with group or individual painphysiology education a successful treatment for chronic low

back pain J Man Manipul Therap 20031188ndash9470 Birklein F Complex regional pain syndrome J Neurol

2005252131ndash871 Weber M Birklein F Neundorfer B Schmelz M Facilitated

neurogenic inflammation in complex regional pain syndromePain 200191251ndash7

72 Huygen F de Bruijn AGJ de Bruin MT Groeneweg JG KleinJ Zijlstra FJ Evidence for local inflammation in complexregional pain syndrome type 1 Mediat Inflamm 20021147ndash51

73 Koban M Leis S Schultze-Mosgau S Birklein F Tissuehypoxia in complex regional pain syndrome Pain2003104149ndash57

74 Janig W Baron R Complex regional pain syndrome mysteryexplained Lancet Neurol 20032687ndash97

75 Janig W Baron R Complex regional pain syndrome is adisease of the central nervous system Clin Autonom Res200212150ndash64

76 Wasner G Heckmann K Maier C Baron R Vascularabnormalities in acute reflex sympathetic dystrophy (CRPS I)complete inhibition of sympathetic nerve activity withrecovery Arch Neurol 199956613ndash20

77 Goebel A Vogel H Caneris O Bajwa Z Clover L Roewer Net al Immune responses to Campylobacter and serumautoantibodies in patients with complex regional painsyndrome J Neuroimmunol 2005162184ndash9

78 Goebel A Stock M Deacon R Sprotte G Vincent AIntravenous immunoglobulin response and evidence forpathogenic antibodies in a case of complex regional painsyndrome 1 Ann Neurol 200557463ndash4

79 Alexander GM van Rijn MA van Hilten JJ Perreault MJSchwartzman RJ Changes in cerebrospinal fluid levels of pro-inflammatory cytokines in CRPS Pain 2005116213ndash9

80 Moseley GL Why do people with CRPS1 take longer torecognise their affected hand Neurology 2004622182ndash6

81 Krause P Forderreuther S Straube A Motor corticalrepresentation in patients with complex regional painsyndrome A TMS study Schmerz 200620181

82 Maihofner C Neundorfer B Birklein F Handwerker HOMislocalization of tactile stimulation in patients with complexregional pain syndrome J Neurol 2006253772ndash9

83 Pleger B Ragert P Schwenkreis P Forster AF Wilimzig CDinse H et al Patterns of cortical reorganization parallelimpaired tactile discrimination and pain intensity in complexregional pain syndrome Neuroimage 200632503ndash10

84 Drummond PD Finch PM Sensory changes in the foreheadof patients with complex regional pain syndrome Pain200612383ndash9

85 McCabe CS Haigh RC Halligan PW Blake DR Referredsensations in patients with complex regional pain syndrometype 1 Rheumatology 2003421067ndash73

86 Moseley GL Distorted body image in complex regional painsyndrome type 1 Neurology 200565773

87 Moseley GL Imagined movements cause pain and swelling ina patient with complex regional pain syndrome Neurology2004621644

88 Acerra N Moseley GL Dysynchiria Watching the mirrorimage of the unaffected limb elicits pain on the affected sideNeurology 200565751ndash3

89 Craig AD The neural representation of the physiologicalcondition of the body pain as an aspect of homeostasis JPhysiol (Lond) 200153616S

90 Grande LA Loeser JD Ozuna J Ashleigh A Samii AComplex regional pain syndrome as a stress response Pain2004110495ndash8

91 van de Beek WJT Roep BO van der Slik AR Giphart MJ vanHilten BJ Susceptibility loci for complex regional pain

RECONCEPTUALISING PAIN ACCORDING TO MODERN PAIN SCIENCE 177

Pub

lishe

d by

Man

ey P

ublis

hing

(c)

W S

Man

ey amp

Son

Lim

ited

syndrome Pain 200310393ndash792 Moseley GL Graded motor imagery for pathologic pain ndash

a randomized controlled trial Neurology 2006672129ndash3493 Moseley GL Graded motor imagery is effective for long-

standing complex regional pain syndrome Pain 2004108192ndash894 Parsons LM Integrating cognitive psychology neurology and

neuroimaging Acta Psychol 2001107155ndash8195 Virtual imagined and mirror movements - a novel approach to

complex regional pain syndrome (CRPS1) EuropeanFederation of IASP Chapters Triennial Conference 2003Prague Czech Republic

96 Lotze M Montoya P Erb M Hulsmann E Flor H Klose U etal Activation of cortical and cerebellar motor areas duringexecuted and imagined hand movements an fMRI study JCognitive Neurosci 199911491ndash501

97 McCabe CS Haigh RC Ring EFJ Halligan PW Wall PDBlake DR A controlled pilot study of the utility of mirrorvisual feedback in the treatment of complex regional painsyndrome (type 1) Rheumatology 20034297ndash101

98 Juottonen K Gockel M Silen T Hurri H Hari R Forss NAltered central sensorimotor processing in patients withcomplex regional pain syndrome Pain 200298315ndash23

99 Eisenberg E Chistyakov AV Yudashkin M Kaplan B HafnerH Feinsod M Evidence for cortical hyperexcitability of theaffected limb representation area in CRPS a psychophysical

and transcranial magnetic stimulation study Pain200511399ndash105

100 Pleger B Tegenthoff M Schwenkreis P Janssen F Ragert PDinse HR et al Mean sustained pain levels are linked tohemispherical side-to-side differences of primary somatosensorycortex in the complex regional pain syndrome I Exp Brain Res2004155115ndash9

101 Flor H Denke C Schaefer M Grusser S Effect of sensorydiscrimination training on cortical reorganisation andphantom limb pain Lancet 20013571763ndash4

102 Acerra N Souvlis T Moseley GL Common findings in strokecomplex regional pain syndrome and phantom limb painImplications and future directions Clin Rehab Med 2006 In press

103 Cortical issues with rehabilitation and learning 8thInternational Congress of the Australian PhysiotherapyAssociation 2004 Adelaide Australia

104 Lamotte RH Campbell JN Comparison of responses ofwarm and nociceptive C-fiber afferents in monkey with humanjudgments of thermal pain J Neurophysiol 197841509ndash28

105 Meyer RA Campbell JN Myelinated nociceptive afferentsaccount for the hyperalgesia that follows a burn to the handScience 19812131527ndash9

106 Meyer DE Kieras DE A computational theory of executivecognitive processes and multiple-task performance Part 1Basic mechanisms Psychol Rev 19971043ndash65

178 MOSELEY

G LORIMER MOSELEY

Nuffield Medical Research Fellow Le Gros Clark Building Department of Physiology Anatomy amp Genetics

Oxford University Oxford OX1 3QX UK

Tel +44 (0)1865 282658 Fax +44 (0)1865 282656 E-mail lorimermoseleymedscioxacuk