Articles

Graded motor imagery for

pathologic pain
A randomized controlled trial
G. Lorimer Moseley, PhD
AbstractBackground: Phantom limb and complex regional pain syndrome type 1 (CRPS1) are characterized by changes
in cortical processing and organization, perceptual disturbances, and poor response to conventional treatments. Graded
motor imagery is effective for a small subset of patients with CRPS1. Objective: To investigate whether graded motor
imagery would reduce pain and disability for a more general CRPS1 population and for people with phantom limb pain.
Methods: Fifty-one patients with phantom limb pain or CRPS1 were randomly allocated to motor imagery, consisting of 2
weeks each of limb laterality recognition, imagined movements, and mirror movements, or to physical therapy and
ongoing medical care. Results: There was a main statistical effect of treatment group, but not diagnostic group, on pain
and function. The mean (95% CI) decrease in pain between pre- and post-treatment (100 mm visual analogue scale) was
23.4 mm (16.2 to 30.4 mm) for the motor imagery group and 10.5 mm (1.9 to 19.2 mm) for the control group. Improvement
in function was similar and gains were maintained at 6-month follow-up. Conclusion: Motor imagery reduced pain and
disability in these patients with complex regional pain syndrome type I or phantom limb pain, but the mechanism, or
mechanisms, of the effect are not clear.
NEUROLOGY 2006;67:21292134
Complex regional pain syndrome type 1 (CRPS) is
considered a pathologic pain syndrome because the
pain does not seem to reflect the underlying tissue
pathology.
1
However, the pathophysiology of CRPS1
is not well understood: peripheral and central
changes have been observed and altered central rep-
resentation of perceptual, motor, and autonomic sys-
tems have been implicated.
1
If such cortical
mechanisms underpin the disease, it would seem
reasonable to target them in treatmenttrain the
brain. One such approach, graded motor imagery,
reduced pain and disability in a relatively homoge-
nous sample of patients with CRPS1 after wrist frac-
ture, all of whom had motor dysfunction as part of
their condition.
2,3
Although those clinical trials ap-
pear encouraging, about 50% of subjects were ex-
cluded, so whether the approach is effective for a
wider CRPS1 population is not known. The first aim
of the current study was to resolve this issue.
Graded motor imagery might also be effective in
those with phantom limb pain, which is also consid-
ered a pathologic pain syndrome
4
and is also thought
to be dominated by altered cortical mechanisms. The
similarities between phantom limb pain and CRPS1,
which have been noted elsewhere,
4-8
suggest that
graded motor imagery may be effective for phantom
limb pain as well as CRPS1. The second aim of the
current study was to determine if this is the case.
Methods. Design. A single blind randomized trial with
6-month follow-up was conducted (figure 1).
Eligible participants were drawn from three patient groups:
patients with phantom limb pain after amputation of one limb,
phantom limb pain (within deafferented zone, according to results
of previous quantitative sensory testing, not verified here) after
Additional material related to this article can be found on the Neurology
Web site. Go to www.neurology.org and scroll down the Table of Con-
tents for the December 26 issue to find the title link for this article.
Editorial, see page 2115
This article was previously published in electronic format as an Expedited E-Pub on November 2, 2006, at www.neurology.org.
From the Department of Physiology, Anatomy & Genetics & fMRIB Centre, University of Oxford, UK; and the School of Physiotherapy, The University of
Sydney, Australia.
G.L.M. is a Nuffield Medical Fellow at Oxford University and is on leave from the School of Physiotherapy, The University of Sydney, Australia.
Disclosure: The author reports no conflicts of interest.
Received December 12, 2005. Accepted in final form August 8, 2006.
Address correspondence and reprint requests to Dr. Lorimer Moseley, Department of Physiology, Anatomy & Genetics, University of Oxford, South Parks
Road, Oxford OX1 3QX, UK; e-mail: [email protected]
Copyright 2006 by AAN Enterprises, Inc. 2129
brachial plexus avulsion injury of one arm, and patients with
complex regional pain syndrome, type 1 (CRPS1). Sixty-nine eligi-
ble patients (37 F) were contacted via hospital physiotherapy de-
partment, neurology, and pain clinic waiting lists. Patients were
excluded if they had been diagnosed with any other neurologic,
psychopathology, or motor disorder or dyslexia; had difficulty per-
forming a rapid naming task; were visually impaired; had any
other limb pathology or pain, or lived outside the immediate met-
ropolitan area of the host department. Nine patients were ex-
cluded according to those criteria. Nine patients with CRPS1 who
did not fulfill recognized diagnostic criteria
9
were also excluded,
which left 51 patients (32 F). This sample size would detect an
effect size of 0.80 (equivalent to a reduction in pain of 29 mm on a
100 mm visual analogue scale), with a probability of 80%, as-
suming 0.05. There was no adjustment for multiple end
points. Written informed consent was obtained and all procedures
were approved by the institutional research ethics committee and
conformed to the Declaration of Helsinki.
Protocol. Patients were randomized via random number gen-
eration by an independent investigator, to a graded motor imagery
program (experimental group) or to standard medical and physio-
therapy care (control group), using a random numbers table. Prior
to randomization, an independent investigator obtained several
assessments.
Assessments. Questionnaires. Patient-specific task-related
numerical rating scale (NRS): patients were asked to select five
activities or tasks that they regularly performed prior to their
injury but now found difficult to perform because of pain. Using
an 11-point numerical rating scale (NRS) anchored with 0, com-
pletely unable to perform and 10, able to perform normally,
they were asked How well can you perform that task now? This
measure is based on similar measures validated in patients with
neck pain
10
and knee pain.
11
It is sensitive to change in people
with upper limb CRPS1 after wrist fracture.
3
In the current study,
this measure estimated functional level and was a primary out-
come variable. It is herein referred to as function NRS.
The McGill Pain Questionnaire (MPQ)
12
: patients completed
the MPQ with regard to current pain, but the visual analogue
scale (VAS) for pain intensity referred to the average level of pain
over the previous 2 days. This measure of pain was a primary
outcome variable and is herein referred to as pain VAS.
Successful response to treatment was estimated in three differ-
ent ways: 1) pain VAS decreased by 50%, 2) function NRS in-
creased 4 points, and 3) pain VAS decreased by 50% and
function NRS increased 4 points. These criteria are consistent
with recommendations made in the pain literature
13
and with
special regard to patients with CRPS1.
14
The number needed treat
(NNT) in order to get a successful response from the motor imag-
ery program that would not be imparted by the control treatment
was calculated for each criterion, for post program and for 6
month follow-up.
Clinical assessment. Diagnostic criteria for CRPS1
9
were as-
sessed before and after the treatment period by independent in-
vestigators blinded to treatment group. The following clinical
findings were also recorded for phantom limb pain patients: symp-
toms and signs of hyperalgesia and allodynia of the stump for
amputees; symptoms of swelling and temperature changes in the
phantom limb; and symptoms of motor disorders of the phantom
limb (patients were asked whether the phantom felt as though it
was cramping, cramped, or stuck in a particular position).
Assessments were undertaken and entered into a datasheet by
an independent investigator who was blind to group and assess-
ment occasion. All assessments were undertaken at prerandom-
ization and at 6 weeks (completion of the treatment period). Pain
VAS and function NRS were also undertaken at 6 months
follow-up.
Active treatmentmotor imagery program. The first 2 weeks
were the limb laterality recognition phase. Forty photographs of a
right hand, matched to gender, and in various positions and align-
ments, were digitally mirrored to create a bank of 80 images.
Twenty-four photographs of a right foot, matched to gender, and
in various positions and alignments, were digitally mirrored to
create a bank of 48 images. Thus, there were four banks of im-
ages: upper and lower limbs, male and female. Every image in the
appropriate bank of images was classified by each subject into one
of four categories, according to the level of pain that would be
provoked if the subject were to adopt the position shown in the
image. In a previous sample of patients with CRPS1, this catego-
rization explained 45% of the variance in RT to recognize the
laterality of the pictured limb [F(1,54) 46.6; adjusted R
2
0.45;
p 0.001] and was used here as an estimate of task difficulty.
15
This estimate of task difficulty was used to increase the training
load for the limb laterality recognition phase of the motor imagery
program according to table 1. An in-house software program and
Figure 1. Trial plan. CRPS complex regional pain syn-
drome; BPAI brachial plexus avulsion injury; NPS
neuropathic pain scale; MPQ McGill Pain question-
naire; MIP motor imagery program.
Table 1 Protocol for training load during each phase of the motor imagery program
Phase Day 14 Day 58 Day 814
Limb laterality recognition Categories 12 (80 trials) Categories 13 (120 trials) Categories 24 (120 trials)
Imagined movements Category 1 (20 trials) Categories 12 (40 trials) Categories 13 (60 trials)
Mirror movements Category 1 (20 trials) Categories 12 (40 trials) Categories 12 (80 trials)
2130 NEUROLOGY 67 December (2 of 2) 2006
laptop computer were used to randomly present images from the
appropriate image bank. Patients responded by pushing the left or
right pressure-sensitive button, according to whether the picture
showed a left or right limb. The software program recorded time of
trial and response time (RT) and accuracy of each response. Cor-
rect responses were analyzed using the ratio between mean RT for
the affected limb and mean RT for the unaffected limb.
The next 2 weeks were the imagined movements phase. Im-
ages of both limbs were randomly presented and subjects were
advised to imagine twice adopting the posture shown with a
smooth and pain-free movement. Subjects were advised not to
imagine watching themselves perform the movement but to imag-
ine actually performing the movement. Training load for imagined
movements was increased according to table 1.
The next 2 weeks was the mirror movements phase, which
used a purpose-built mirror box (single aperture 300 mm 300
mm; cardboard with Perspex mirror on one external surface)
(NOIgroup.com, Adelaide, Australia). Subjects were advised to
twice adopt the posture shown with both hands, using smooth and
pain-free movements. Training load for mirror movements was
increased according to table 1. This clinical regimen is similar to
one that has been discussed in detail previously.
2,3
Throughout the motor imagery program, there were weekly
consultations with the physiotherapist who supervised perfor-
mance, monitored progress, and answered any questions raised by
the patient. Because this study aimed to compare this treatment
to standard care, patients were advised not to participate in other
treatments during the study period. The software program re-
corded participation at home overtly during the first two phases
and overtly by a training diary during the last phase.
Control treatmentmedical/physiotherapy management.
Subjects allocated to the control group undertook a 6-week physio-
therapy treatment program and maintained usual medical care.
Physiotherapists were instructed not to include treatments that
were similar to those used in the experimental group, or that used
mirrors or imagined movements. Physiotherapists were advised to
include at least one treatment per week and a home program that
involved a training load comparable to that in the motor imagery
program (i.e., hourly training). Participation was recorded via a
training diary.
Subjects in both groups were advised not to commence new
treatments or medications, nor to reduce other treatments or med-
ications, unless they were instructed by their treating clinician to
do so. There were no instructions about management during the
follow-up period, although at 6 months, subjects were asked what
treatments they had received during that period.
Statistical analysis. All statistics were performed using SPSS
11.0.0 (SPSS, Chicago, IL). To compare mean response between
treatment groups at 6 weeks or at 6 months, an analysis of covari-
ance model (ANCOVA) was constructed, with covariates pretreat-
ment score and diagnosis. The primary outcome was pain VAS or
function NRS at 6 weeks or 6-month follow-up (ordinal data).
Diagnosis was clinical group, dummy-coded as 0 for phantom limb
pain (BPAI or amputee patients) and 1 for CRPS1.
Because one proposed mechanism of effect of graded motor
imagery involves normalization of cortical organization, which in
turn is related to duration of symptoms,
16
a secondary analysis
investigated the relationship between the duration of symptoms
and the response to treatment, via two linear regressions between
duration and change in pain VAS at post treatment and at
follow-up.
For all analyses, significance was set at 0.05.
Results. Subjects. All data were collected over 32
months. One female subject in the control group withdrew
from the study because she sustained an unrelated injury.
There were no other dropouts or withdrawals. Subject
characteristics according to group are shown in table E-1
on the Neurology Web site at www.neurology.org. There
was no difference between the groups by gender (p 0.56
by
2
test), age, duration of disease, VAS pain score, or
NRS score (p 0.50 for all, by two-sample t-tests).
Pain and function at post-program. Statistical results
are described here, NNTs in table 2 and results for specific
diagnoses are presented in figure 2. The regressions
showed effects for pain VAS [ANCOVA F(3,46) 6.77, p
0.001] and for function NRS [ANCOVA F(3,46) 8.95, p
0.001]. For pain intensity at post-program, there was a
Table 2 Mean (95% CI) for number needed to treat (NNT) to achieve a preset change in pain or function or both
NNT to get a
50% decrease in pain
NNT to get a 4-point
increase in function
NNT for
both criteria
Response at post-program 3 (26) 4 (211) 4 (217)
Response at 6-mo. follow-up 2 (15) 2 (15) 3 (24)
Pain average pain intensity over previous 2 days; Function ability to perform five patient-selected activities.
Figure 2. Mean (columns) and SD (error bars) change in
(A) average pain level over the past 2 days, measured with
a 100 mm visual analogue scale (VAS), and (B) average
score for each of five patient-selected tasks, on an 11-point
numerical rating scale (NRS), for patients in the control
group and the motor imagery group. Data are shown for
patients with complex regional pain syndrome (CRPS1,
open columns), phantom limb pain after brachial plexus
avulsion injury (BPAI, vertical stripes) or amputation
(Amp, filled columns), and reflect the change between pre-
and post-treatment (left column of each pair) and between
pretreatment and 6-month follow-up (right column of each
pair). No change is marked by the horizontal dashed line
at zero.
December (2 of 2) 2006 NEUROLOGY 67 2131
main effect of treatment group (unstandardized B
1.298, p 0.002). That is, the mean (95% CI) decrease in
average pain over the last 2 days, as measured on the 100
mm VAS, was 23.4 mm (16.2 to 30.4 mm) for the MIP
group and 10.5 mm (1.9 to 19.2 mm) for the control group.
For function NRS, there was a main effect of treatment
group (unstandardized B 1.532, p 0.001). That is, the
mean (95% CI) increase in average score for each patient-
specific task, measured using a 0 to 10 NRS, was 2.2 points
(1.33.0 points) for the MIP group and 0.6 points (0.21.0
points) for the control group.
Pain and function at 6-month follow-up. The regres-
sions showed effects for pain VAS [ANCOVA F(3,46)
8.701, p 0.001] and for function NRS [ANCOVA
F(3,46) 7.327, p 0.001]. For pain VAS at follow-up,
there was a main effect of treatment group (unstandard-
ized B 2.07, p 0.001) (figure 2B). That is, the mean
(95% CI) decrease in pain VAS between pretreatment and
follow-up was 32.1 mm (23.840.3 mm) for the MIP group
and 11.6 mm (2.420.7 mm) for the control group. The
effect size at 6-month follow-up, but not at post-program,
was consistent with the estimated effect size.
For function NRS, there was a main effect of treatment
group (unstandardized B 2.18, p 0.001) (figure 2B).
That is, the mean (95% CI) increase in task-specific NRS
VAS between pretreatment and follow-up was 3.7 points
(2.74.6 points) for the MIP group and 1.5 points (12.2
points) for the control group.
Outcome and response to treatment at follow-up: Pain
VAS 3, function NRS 5, or both. NNT was calculated
for pain VAS 3, function NRS 5, and for both, and are
shown, with 95% CI, in table 2. The duration of symptoms
did not relate to change in pain VAS at post-treatment
(p 0.224), nor to change in pain VAS at follow-up (p
0.071).
Participation in home program. Participation with the
home program was 75% throughout the treatment period.
There was no difference between treatment groups in any
phase (p 0.211 for all).
Treatment during follow-up period. Only 11 patients
in the treated group, but all patients in the control group,
reported that they sought treatment for their pain during
the follow-up period (p 0.001 by
2
test). For the treated
group, this constituted between 3 and 11 (median 6)
physiotherapy treatments, including motor imagery, tac-
tile discrimination training, and functional exposure (9 pa-
tients), a multidisciplinary pain management program (1
patient), and between 2 and 4 general practitioner visits (3
patients). For the control group, treatments sought were
physiotherapy treatment (motor imagery, desensitization,
tactile discrimination, functional exposure) for 14 patients
(median number of treatments 12, range 1 to 18), multi-
disciplinary pain management program for 7 patients, gen-
eral practitioner visits for 10 patients (median number of
visits 6, range 2 to 1), spinal cord stimulator for 1
patient.
Discussion. Graded motor imagery reduced pain
and disability in a wider CRPS1 population and in
those with phantom limb pain after amputation or
brachial plexus avulsion injury. The following results
support this position: 1) a significant effect of treat-
ment group on both primary outcome measures (pain
and function), such that pain decreased and function
increased for the motor imagery group, relative to
the control group; 2) NNTs for response to treatment
at 6 months of about 3.
Graded motor imagery has been shown to reduce
pain and disability in a relatively homogenous group
of patients with chronic CRPS1
2,3
and mirror therapy
alone has shown efficacy for those with acute
CRPS1.
17
The current data corroborate those studies,
although the mean magnitude of pain reduction was
about 50% less in the current work than it was in the
earlier studies using graded motor imagery. A likely
contributor to the reduced mean effect is the relative
heterogeneity of the current sample. First, phantom
limb pain and CRPS1, although both categorized
here and elsewhere
4,6,18
as pathologic pain disorders,
and although changes in cortical organization and
some clinical findings are common to both, are fun-
damentally different: whereas CRPS usually occurs
after minor injury and involves no demonstrable
nerve injury, phantom limb pain occurs after major
trauma and undeniable nerve injury. Further to
that, phantom pain after brachial plexus avulsion
injury may involve different mechanisms to phantom
pain after amputation. That is, the current design
may conceal stronger effects in one group than an-
other and that the effect occurred regardless of diag-
nostic group does not imply that the same
mechanisms underpin that effect in each group. This
study was underpowered to systematically evaluate
different response profiles between the diagnostic
groups.
Even the current sample of patients with CRPS1
was more heterogeneous than previous samples: pre-
vious studies limited inclusion to CRPS1 initiated by
wrist fracture, limited disease duration to more than
6 months, and excluded those without motor signs
and symptoms. The final issue is particularly impor-
tant because diagnosis of CRPS1 is based on present-
ing signs and symptoms, not on mechanisms, and
many contributing mechanisms exist.
1
Perhaps sub-
classes of CRPS1 exist and some are better suited to
motor imagery training than others.
The control group here was intended to reflect
current practice, which is dominated by drug and
physical therapies, implemented according to a
cognitive-behavioral model.
19-21
That the medical
component of the control group treatment consti-
tuted ongoing medical care and therefore was not
new probably reduced its effect because most treat-
ment is effective to some extent when it is first start-
ed.
22
Importantly, however, the commencement of
conventional physiotherapy in addition to ongoing
medical care had no discernible effect on pain or
function. This seems discouraging, although the cur-
rent study limited the duration and frequency of
physical therapy, which suggests that firm conclu-
sions are probably premature.
Although the patients used here were broadly
similar to those involved in other studies
19,23-27
in-
cluding therapeutic trials,
17
it is not possible to di-
2132 NEUROLOGY 67 December (2 of 2) 2006
rectly compare the current results to established
data from CRPS1 or phantom limb pain because ro-
bust trials of other therapies are lacking. This lack of
data probably explains why tricyclic antidepres-
sants, antiepileptics, and opioids remain the main-
stay of treatment, particularly with phantom limb
painthese drugs have demonstrated some efficacy
in other types of neuropathic pain.
28
That said, the
demonstrated effect of such drugs in those groups is
not better than the current results. For example, a
systematic review of drug therapies for peripheral
neuropathic pain reported NNTs of between 2.2 and
5.0 for tricyclics, antiepileptics, and opioids, and be-
tween 2.9 and 7.7 for selective serotonin reuptake
inhibitors.
29
The present study found NNTs of about
3 for similar outcome parameters.
How progressive motor imagery reduced pain and
disability is not known. This study was based on the
idea of applying established principles of rehabilita-
tion to training the brain. That limb laterality recog-
nition activates premotor cortices but not primary
motor cortex, whereas imagined movements activate
both,
30
and that the order of hand laterality recogni-
tion, imagined movements, and mirror movements
seems to be important in the effect motor imagery,
3
seem consistent with this possibility. Perhaps prac-
ticing laterality recognition is to limb movement
31
as
practicing knee movement is to walking. This seems
sensible in principle, but the neural mechanisms by
which it might occur are unknown.
An alternative possible explanation for the effect
of motor imagery is that it promotes sustained atten-
tion to the affected limb. That probably does not
explain the effect in patients with CRPS1 post-wrist
fracture, because undertaking the treatment compo-
nents in a different order does not work,
3
but it may
explain the effect observed here. Such an explana-
tion would seem consistent with several findings in
the literature. For example, patients with CRPS1
15,32
and with phantom limb pain
33
take longer to recog-
nize the laterality of a pictured limb if it corresponds
to their own affected limb. Acute experimental
pain,
34
and the expectation of acute experimental
pain,
35
delays recognition of the laterality of the op-
posite limb, which implies an information processing
bias toward the body part in pain, yet patients with
CRPS1 and phantom limb pain show the opposite
effect, which implies an opposite information pro-
cessing biasaway from the body part in pain. If so,
progressive motor imagery may simply serve to re-
verse that tendency. That speculation remains to be
verified, but other signs and symptoms in CRPS1
and after amputation or brachial plexus avulsion are
consistent with some sort of neglect: symptoms of
foreignness and neglect
25,36
; the perception that their
affected limb is bigger than it is,
37
which also occurs
after anesthesia.
38
The question, then, is why would a neglect-like
disorder cause pain? Pain is not characteristic of the
usual parietal neglect syndromes. The cortical model
of pathologic pain
4
is relevant here because it pro-
poses that the changes in cortical proprioceptive rep-
resentation, themselves related to the neglect-like
phenomena outlined above, underpin the pain be-
cause they evoke an incongruence between motor
commands and sensory feedback. It is established
that 1) cortical proprioceptive representation is al-
tered in CRPS1 and in phantom limb pain,
39-41
2)
that the extent of reorganization relates to the dura-
tion of symptoms
42
(it is notable that although the
current study did not find a relationship between
duration of symptoms and response to motor imag-
ery, it was probably underpowered [p 0.071] to do
so), and 3) that resolution of symptoms is associated
with normalization of cortical reorganization.
43
There is a large amount of literature on sensory-
motor incongruence, typified by the reafference prin-
ciple, whereby an exact copy of the command for
movement is subtracted from sensory input about
the actual movement to yield an error signal.
44
How-
ever, only recently has sensory-motor incongruence
been empirically investigated with regard to pain
45,46
and there are insufficient data to conclude whether
motor imagery might remove sensory-motor incon-
gruence. If it does, this mechanism may also mediate
the effect of sensory discrimination training in phan-
tom limb pain
24
and of sensory-motor retuning in
CRPS1, for which there is preliminary evidence.
43
Alternatively, perhaps the neglect-like phenomena
and the pain are not causally related but both result
from, for example, altered central representations of
somatic input in the thalamus or cortex,
1
in which
case motor imagery might primarily serve to normal-
ize these central representations. Thus, although ev-
idence is emerging that treatments such as graded
motor imagery and sensory discrimination training
can be effective for pathologic pain, further studies
are needed to replicate the current data and eluci-
date the mechanisms involved.
Acknowledgment
The author thanks Emily Briars, Claire Lee, and Owen Parker for
their contribution to treatment and data collection.
References
1. Janig W, Baron R. Complex regional pain syndrome: mystery ex-
plained? Lancet Neurol 2003;2:687697.
2. Moseley GL. Graded motor imagery is effective for long-standing com-
plex regional pain syndrome. Pain 2004;108:192198.
3. Moseley GL. Is successful rehabilitation of complex regional pain syn-
drome due to sustained attention to the affected limb? A randomised
clinical trial. Pain 2005;114:5461.
4. Harris AJ. Cortical origin of pathological pain. Lancet 1999;354:1464
1466.
5. Acerra N, Moseley GL. Dysynchiria: watching the mirror image of the
unaffected limb elicits pain on the affected side. Neurology 2005;65:
751753.
6. McCabe CS, Haigh RC, Halligan PW, Blake DR. Referred sensations in
patients with complex regional pain syndrome type 1. Rheumatology
2003;42:10671073.
7. McCabe CS, Moseley GL. Functional strategies of restoration of com-
plex regional pain syndrome. In: Justins D, ed. Refresher course
syllabus (pain 2005an updated review). Amsterdam: IASP Press,
2005.
8. Moseley GL. Making sense of S1 maniaare things really that simple?
In: Gifford L, ed. Topical Issues in Pain Volume 5. Falmouth: CNS
Press, 2006:321340.
December (2 of 2) 2006 NEUROLOGY 67 2133
9. Bruehl S, Harden RN, Galer BS, et al. External validation of IASP
diagnostic criteria for complex regional pain syndrome and proposed
research diagnostic criteria. Pain 1999;81:147154.
10. Westaway MD, Stratford PW, Binkley JM. The patient-specific func-
tional scale: validation of its use in persons with neck dysfunction.
J Orthop Sports Phys Ther 1998;27:331338.
11. Chatman AB, Hyams SP, Neel JM, et al. The Patient-Specific Func-
tional Scale: measurement properties in patients with knee dysfunc-
tion. Phys Ther 1997;77:820829.
12. Melzack R. The short-form McGill Pain Questionnaire. Pain 1987;30:
191197.
13. McQuay HJ, Moore RA, Eccleston C, Morley S, Williams AC. System-
atic review of outpatient services for chronic pain control. Health Tech-
nol Assess 1997;1:iiv, 1135.
14. Forouzanfar T, Weber WEJ, Kemler M, van Kleef M. What is a mean-
ingful pain reduction in patients with complex regional pain syndrome
type 1? Clin J Pain 2003;19:281285.
15. Moseley GL. Why do people with CRPS1 take longer to recognise their
affected hand? Neurology 2004;62:21822186.
16. Flor H. The functional organization of the brain in chronic pain. Prog
Brain Res 2000;129:313322.
17. McCabe CS, Haigh RC, Ring EFJ, Halligan PW, Wall PD, Blake DR. A
controlled pilot study of the utility of mirror visual feedback in the
treatment of complex regional pain syndrome (type 1). Rheumatology
2003;42:97101.
18. Alexander GM, van Rijn MA, van Hilten JJ, Perreault MJ, Schwartz-
man RJ. Changes in cerebrospinal fluid levels of pro-inflammatory cy-
tokines in CRPS. Pain 2005;116:213219.
19. Forouzanfar T, Koke AJA, van Kleef M, Weber WEJ. Treatment of
complex regional pain syndrome type I. Eur J Pain 2002;6:105122.
20. Wasner G, Schattschneider J, Binder A, Siebrecht D, Maier C, Baron R.
Recent trends in understanding and therapy of complex regional pain
syndromes. Anaesthesist 2003;52:883895.
21. McBride A, Atkins R. Complex regional pain syndrome. Curr Orthopae-
dics 2005;19:155165.
22. Moerman DE, Jonas WB. Deconstructing the placebo effect and finding
the meaning response. Ann Intern Med 2002;136:471476.
23. Juottonen K, Gockel M, Silen T, Hurri H, Hari R, Forss N. Altered
central sensorimotor processing in patients with complex regional pain
syndrome. Pain 2002;98:315323.
24. Flor H, Denke C, Schaefer M, Grusser S. Effect of sensory discrimina-
tion training on cortical reorganisation and phantom limb pain. Lancet
2001;357:17631764.
25. Forderreuther S, Sailer U, Straube A. Impaired self-perception of the hand
in complex regional pain syndrome (CRPS). Pain 2004;110:756761.
26. MacLachlan M, McDonald D, Waloch J. Mirror treatment of lower limb
phantom pain: a case study. Disabil Rehabil 2004;26:901904.
27. Srinivasan AV, Velmurugendran UC, Nadu T, Ramachandran RD,
Ramachandran VS. Allesthesia and extinction of referred sensations
after brachial plexus avulsion. Neurology 1998;50:A256A256.
28. Birklein F. Complex regional pain syndrome. J Neurology 2005;252:
131138.
29. Beniczky S, Tajti J, Varga ET, Vecsei L. Evidence-based pharmacologi-
cal treatment of neuropathic pain syndromes. J Neural Transm 2005;
112:735749.
30. Moseley GL, Schweinhardt P, Wise R, Tracey I. Virtual, imagined and
mirror movementsa novel approach to complex regional pain syn-
drome (CRPS1). In: Europ Fed IASP Chapt Triennial Conference; Pra-
gue, Czech Republic; 2003.
31. Parsons LM. Integrating cognitive psychology, neurology and neuroim-
aging. Acta Psychol 2001;107:155181.
32. Schwoebel J, Coslett HB, Bradt J, Friedman R, Dileo C. Pain and the
body schema: effects of pain severity on mental representations of
movement. Neurology 2002;59:775777.
33. Nico D, Daprati E, Rigal F, Parsons L, Sirigu A. Left and right hand
recognition in upper limb amputees. Brain 2004;127:120132.
34. Moseley GL, Sim DF, Henry ML, Souvlis T. Experimental hand pain
delays recognition of the contralateral handevidence that acute and
chronic pain have opposite effects on information processing? Cogn
Brain Res 2005;25:188194.
35. Hudson ML, McCormick K, Zalucki N, Moseley GL. Expectation of pain
replicates the effect of pain in a hand laterality recognition task: bias in
information processing toward the painful side? Eur J Pain 2006;10:
219224.
36. Galer BS, Jensen M. Neglect-like symptoms in complex regional pain
syndrome: results of a self-administered survey. J Pain Symptom Man-
age 1999;18:213217.
37. Moseley GL. Distorted body image in complex regional pain syndrome
type 1. Neurology 2005;65:773.
38. Gandevia S, Phegan C. Perceptual distortions of the human body image
produced by local anaesthesia, pain and cutaneous stimulation. J Phys
(Lond) 1999;514:609616.
39. Flor H, Elbert T, Knecht S, et al. Phantom-limb pain as a perceptual
correlate of cortical reorganization following arm amputation. Nature
1995;375:482484.
40. Maihofner C, Handwerker HO, Neundorfer B, Birklein F. Patterns of
cortical reorganization in complex regional pain syndrome. Neurology
2003;61:17071715.
41. Pleger B, Tegenthoff M, Schwenkreis P, et al. Mean sustained pain
levels are linked to hemispherical side-to-side differences of primary
somatosensory cortex in the complex regional pain syndrome I. Exp
Brain Res 2004;155:115119.
42. Flor H. Cortical reorganisation and chronic pain: implications for reha-
bilitation. J Rehabil Med 2003;35:6672.
43. Pleger B, Tegenthoff M, Ragert P, et al. Sensorimotor returning in
complex regional pain syndrome parallels pain reduction. Ann Neurol
2005;57:425429.
44. Gandevia S. Kinesthesia: roles for afferent signals and motor com-
mands. In: Rothwell L, Shepherd J, eds. Handbook of physiology, sec-
tion 12, exercise: regulation and integration of multiple systems. New
York: Oxford University Press, 1996:128172.
45. McCabe CS, Haigh RC, Halligan PW, Blake DR. Simulating sensory-
motor incongruence in healthy volunteers: implications for a cortical
model of pain. Rheumatology 2005;44:509516.
46. Moseley GL, McCormick K, Hudson M, Zalucki N. Disrupted cortical
proprioceptive representation evokes symptoms of peculiarity,
foreignness and swelling, but not pain. Rheumatology 2006;45:196
200.
ALERT: NEUROLOGY NOW USING ONLINE PEER REVIEW AND
MANUSCRIPT SUBMISSION SYSTEM
Neurology is now using an online peer review and manuscript submission system called BenchPress.
Authors should upload all original submissions via the Neurology website (www.submit.neurology.org). The
Instructions to Authors detail the submission process and adjusted specifications.
2134 NEUROLOGY 67 December (2 of 2) 2006