NeuroImage 130 (2016) 134144

Contents lists available at ScienceDirect

NeuroImage

journal homepage: www.elsevier.com/locate/ynimg

Full Length Articles

Primary motor cortex changes after amputation correlate with phantom

limb pain and the ability to move the phantom limb
Estelle Rafn a,b,c,d,, Nathalie Richard e,f, Pascal Giraux a,, Karen T. Reilly f,g,,1
a
Department of Physical Medicine and Rehabilitation, Faculty of Medicine, University Jean Monnet of Saint Etienne, F-42023 Saint-Etienne, France
b
University Grenoble Alpes, Grenoble Institut des Neurosciences, GIN, F-38000 Grenoble, France
c
Inserm, U1216, F-38000 Grenoble, France
d
Danish Research Centre for Magnetic Resonance, Centre for Functional and Diagnostic Imaging and Research, Copenhagen University Hospital Hvidovre, Denmark
e
Center for Cognitive Neuroscience, UMR 5229, CNRS, Lyon, France
f
University Claude Bernard Lyon 1, F-69000 Lyon, France
g
INSERM U1028, CNRS UMR5292, Lyon Neuroscience Research Centre, ImpAct Team, F-69000 Lyon, France

a r t i c l e i n f o a b s t r a c t

Article history: A substantial body of evidence documents massive reorganization of primary sensory and motor cortices follow-
Received 26 June 2015 ing hand amputation, the extent of which is correlated with phantom limb pain. Many therapies for phantom
Accepted 15 January 2016 limb pain are based upon the idea that plastic changes after amputation are maladaptive and attempt to normal-
Available online 11 February 2016
ize representations of cortical areas adjacent to the hand area. Recent data suggest, however, that higher levels of
phantom pain are associated with stronger local activity and more structural integrity in the missing hand area
Keywords:
Motor cortex
rather than with reorganization of neighbouring body parts. While these models appear to be mutually exclusive
Reorganization they could co-exist, and one reason for the apparent discrepancy between them might be that no single study has
Motor control examined the organisation of lip, elbow, and hand movements in the same participants. In this study we
Phantom pain thoroughly examined the 3D anatomy of the central sulcus and BOLD responses during movements of the
fMRI hand, elbow, and lips using MRI techniques in 11 upper-limb amputees and 17 healthy control subjects. We ob-
Phantom movements served different reorganizational patterns for all three body parts as the former hand area showed few signs of
reorganization, but the lip and elbow representations reorganized and shifted towards the hand area. We also
found that poorer voluntary control and higher levels of pain in the phantom limb were powerful drivers of
the lip and elbow topological changes. In addition to providing further support for the maladaptative plasticity
model, we demonstrate for the rst time that motor capacities of the phantom limb correlate with post-
amputation reorganization, and that this reorganization is not limited to the face and hand representations but
also includes the proximal upper-limb.
2016 Elsevier Inc. All rights reserved.

Introduction investigating functional reorganization in hand amputees report an

After amputation, the human primary sensorimotor cortex (SM1) un-
dergoes massive reorganization involving both structural (Jiang et al.,
2015; Preissler et al., 2012; Simoes et al., 2012) and functional (Elbert
et al., 1997; Karl et al., 2001; Kew et al., 1994; Knecht et al., 1996;
Montoya et al., 1998; Yang et al., 1994) changes in SM1 contralateral to
the amputation as input-deprived regions are recruited by body-parts
with neighbouring cortical representations. Nearly all studies
Correspondence to: E. Rafn, INSERM U836, Team 11, Grenoble Institute of
Neurosciences, Chemin Fortun Ferrini, Bt EJ Safra, 38700 La Tronche, France.
Correspondence to: K. T. Reilly, INSERM U1028, CNRS UMR5292, Lyon Neuroscience
Research Centre, ImpAct Team, 16, avenue Doyen Lpine, 69500 Bron, France.
E-mail addresses: estelle.rafn@ujf-grenoble.fr (E. Rafn), karen.reilly@inserm.fr
(K.T. Reilly).
1
These authors contributed equally to this work.
inter-hemispheric asymmetry of the face representation (e.g. Karl et al.,
2001; Makin et al., 2015b), and the majority also describes a signicant
correlation between this asymmetry and the degree of phantom limb
pain (Flor et al., 1995; Flor et al., 1998; Karl et al., 2004; fMRI: Lotze
et al., 2001; Lotze et al., 1999; MacIver et al., 2008). This correlation
supports the maladaptive plasticity model, which postulates that
representations of body parts adjacent to the missing limb's representa-
tion expand and invade the deprived cortex, and that this invasion
leads to phantom limb pain. This model is further supported by evidence
that interventions that reduce phantom limb pain also reduce the inter-
hemispheric asymmetry of the face representation (Birbaumer et al.,
1997; Chen et al., 2013; Flor, 2002; Flor et al., 2001; Huse et al., 2001;
Lotze et al., 1999; MacIver et al., 2008).
The maladaptive plasticity model ignores the persistence of the
missing limb's representation (MacIver et al., 2008; Mercier et al.,

http://dx.doi.org/10.1016/j.neuroimage.2016.01.063
1053-8119/ 2016 Elsevier Inc. All rights reserved.
 E. Rafn et al. / NeuroImage 130 (2016) 134144 135

2006; Rafn et al., 2012a), however, and assumes a take-over by other
body parts rather than a cohabitation (see Reilly and Sirigu, 2008 for a
discussion). On the basis of structural and functional MRI data, Makin
and colleagues proposed an alternative model: the persistent representa-
tion model. This model postulates that persistent pain is associated with
preserved structure and function in the former hand area rather than re-
organization of neighbouring body parts (Makin et al., 2013b). These
authors found that individuals with more intense phantom limb pain
had well preserved structure in the former hand territory and greater
activation intensity during phantom movements, as well as decreased
resting state connectivity between the bilateral sensorimotor hand re-
gions. In a more recent paper, the same authors examined lip activations
only. They did observe some form of lip reorganization, in contrast to
previous studies, it was uncorrelated with pain. Instead it was most re-
liably predicted by tactile acuity and dexterity of the intact hand (Makin
et al., 2015b).
While there is empirical evidence in support of both models, neither
provides a complete explanation of post-amputation sensorimotor re-
organization and its relationship with phantom pain and the structure
and function of the former hand territory. Furthermore, both models
ignore the complex relationship between phantom mobility and phan-
tom pain (Gagn et al., 2009; Rafn et al., 2012b) and neither attempts
to explain why the majority of non-pharmaceutical therapies for phan-
tom limb pain try to strengthen the missing limb's representation by
training voluntary or imagined phantom movements (Brunelli et al.,
2015; Foell et al., 2014; Giraux and Sirigu, 2003; MacIver et al., 2008;
Mercier and Sirigu, 2009; Trojan et al., 2014).
A recent modelling study suggests that these two models could co-
exist in the primary somatosensory cortex (Bostrom et al., 2014).
Here, we investigated their possible co-existence in the primary motor
cortex by examining reorganizational patterns of both amputated
body parts (which are informative about representational plasticity)
and non-amputated body parts (which are informative about maladap-
tive plasticity). We performed both structural imaging and functional
recordings by measuring BOLD responses associated with three differ-
ent movements: lips, hands, and elbows. Furthermore, given the clinical
importance of the ability to voluntarily move the phantom limb we
quantied the individual phantom limb motor performance as well as
chronic phantom pain and correlated them with reorganizational
changes of each of the three body parts.

Materials and methods

Participants

Eleven traumatic upper-limb amputees participated in the experiment (mean age 42 years [range: 2267], 4 females and 7 males). On average,
the accident occurred 7.5 years [5 months30 years] before testing. The Edinburgh Handedness Inventory (EHI) revealed that eleven amputees were
right hand dominant prior to the amputation, and ve of these lost their dominant hand (see Table 1 for patient details). Seventeen healthy volun-
teers (7 females and 10 males; mean age 32.1 years [2142], one left-handed person) were recruited from the general population. A detailed medical
history together with careful examination of the anatomical scans conrmed that neither amputees nor controls had suffered a brain lesion or had a
history of neurological or psychiatric illness. The nature of the experimental procedures was explained to all subjects who gave their written in-
formed consent prior to participating in the experiment. This study was approved by the Local Ethics Committee of Lyon Sud-Est IV (A 09-115),
and conformed to the ethical aspects of the Declaration of Helsinki.

Assessment of phantom limb pain and phantom limb motor performance

Clinical data related to the amputation were collected prior to the neuroimaging exam using a structured interview which included an evaluation
of phantom sensations and chronic phantom limb pain intensity perceived over the last three months (PLP), and rated on a continuous 10 cm visual-
to-analog scale (VAS) (Flor et al., 1995). During this interview we also used a standard technique to perform a quantitative assessment of their ability
to make voluntary movements with the phantom limb and of EMG activity associated with these phantom movements (see Gagn et al., 2009 for
details). Patients were asked to produce cyclic voluntary movements with their phantom and to verbally report when they had completed ve move-
ment cycles. EMG activity in the stump muscles veried that amputees were not performing motor imagery (see Rafn et al., 2011, 2012a) and was
used to check the movement times recorded from the verbal start and stop signals. All amputees were able to open and close their phantom hand and
to oppose each of the four ngertips to the thumb. Since hand opening and closing was used during the imaging session to assess motor cortex
activity we quantied phantom motor performance (MP) as the speed of nger-to-thumb opposition (ve divided by the time taken to complete
the ve cycles), with larger MP values associated with better performance.

Table 1
Clinical and phantom limb motor control characteristics of the amputee group. Patients are presented in order of decreasing phantom motor performance.

Age Time since amp. Amp. side/Dom. Amp. level/length of remaining limb Prosthesis use Chronic PLP (VAS) Phantom motor perf.
(years) (months) Side (%) (010) (cycles/s)

33 84 R/R D/80 No use 0 0.34

22 27 L/L P/25 No use 4 0.25
39 360 L/R D/43 No use 0 0.24
28 204 R/R D/48 No use 4 0.22
40 102 R/R P/24 Myo (constant) 6 0.21
29 24 L/R P/25 No use 5 0.14
27 12 L/R P/42 Mechanic (part-time) 3 0.12
67 37 L/R P/32 Esthetic (rarely) 5 0.11
55 106 L/R P/15 Esthetic (rarely) 6 0.11
65 5 L/R P/24 No use 5 0.10
62 24 L/R P/15 Esthetic (constant) 7 0.09

Amp: amputation/amputated; Dom: Dominant; Myo: Myoelectric; L: Left; R: Right; D: Distal; P: Proximal.
136 E. Rafn et al. / NeuroImage 130 (2016) 134144
Morphometric analysis

All subjects (amputees and controls) were scanned using a research-dedicated 1.5 T MRI scanner (Siemens Magnetom Sonata, Erlangen,
Germany) with an 8-channel head coil. Prior to functional protocol acquisitions, high-resolution three-dimensional anatomical images of the
whole brain were acquired (gradient-echo inversion recovery, repetition time 1600 ms, echo time 5 ms, matrix 256 256, eld of view
220 220 mm2). The central sulcus (CS) of each hemisphere of subject was extracted from the MRI data using the Morphologist toolbox of BrainVisa
(http://brainvisa.info) (Perrot et al., 2011; Sun et al., 2011). ER conrmed that the software correctly identied the central sulcus. To control for the
inuence of variable brain size, the CS representations were all normalized to the standard Talairach reference frame (by a 9 parameter afne trans-
formation). Investigation of the 3D sulcal anatomy, the hemispheric comparison and the group comparison of the CS were based on the following
morphological features: surface area (mm2), depth (mm) and thickness (mm).

fMRI paradigm

All subjects were scanned with the head immobilized and eyes open, while performing simple movements with various body parts. The control
group performed ve different movements: opening/closing the ngers and the thumb on the right and left hand, exion/extension of the right and
left elbow, and moving the lips as if blowing a kiss. All amputees performed the lip movement and the hand and elbow movements with their intact
arm. If their amputation was below the elbow (n = 3) they also exed/extended the elbow on the amputated side and opened/closed the phantom
hand. If their amputation was above the elbow (n = 8) the elbow exion/extension and hand opening/closing movements were both performed
with the phantom. Throughout the text the hemispheres are labelled according to the side of the body from which they predominantly process sen-
sorimotor information. Thus, M1-cPHANTOM and M1-cINTACT refer to the hemisphere contralateral to the phantom and intact limbs, respectively. Like-
wise, M1-cDOM and M1-cNON-DOM refer to the hemisphere contralateral to the dominant and non-dominant hands, respectively.
To avoid possible artefacts induced by planning multiple movements, each movement was assessed in a separate fMRI run using a block design.
Each run consisted of 30 movement blocks (block length 10 s) interleaved with rest periods (baselines) of between 5 and 15 s. Auditory beeps pre-
sented throughout the block gave subjects a pacing signal to ensure they moved at a relatively constant rate. For controls, the movement rate during
each block was xed at 0.5 Hz for all body parts. For amputees, lip and elbow movements (in the three below-elbow amputees only) were also per-
formed at 0.5 Hz. For the other movements (hand opening/closing in all amputees and elbow exion/extension in the eight above-elbow amputees),
each amputee chose a pace at which they experienced no additional pain when moving their phantom and could comfortably move the phantom for
10 s. All amputees were able to voluntarily open and close their phantom hand (and ex and extend their phantom elbow for the eight above-elbow
amputees) in time to the auditory pacing signal provided in the scanner. Some amputees reported that movement amplitudes with the phantom
were sometimes smaller than with the corresponding intact side body part.
Movement rate for the phantom movements was established independently for hand and elbow movements before the start of the experiment,
and the same rate was used for both phantom and intact side movements (range: 5 movements in 10 s (0.5 Hz) to one movement in 10 s (0.1 Hz)).
This ensured that phantom movements did not differ from each other or from intact body part movements in terms of effort, and was done to avoid
effort-related differences in motor cortical activation levels. Before each run subjects received verbal instructions about the movement to perform in
that run, and the order of movements was completely randomized for all subjects. At the beginning and end of each block visual start and stop signals
were presented to the subject on a screen at the end of the MRI tunnel that subjects viewed through a 45 oriented mirror placed above the eyes.
Subjects could see this screen throughout the experiment and had no vision of their limbs.
An experimenter sat in the scanner with all subjects in order to ensure that the correct movement was performed for a given acquisition block and
that all movements were unilateral. To ensure amputees performed voluntary phantom movements and not imagined movements (Rafn et al.,
2012a) the experimenter also visually veried the presence of stump muscle contractions during the acquisitions, and in nine of the eleven amputees
surface EMG was recorded from two stump muscles using an MRI-compatible recording system (BIOPAC Systems Inc., Goleta, CA, USA). These data
were visually inspected for the presence of muscle contractions during the active blocks to ensure that patients performed the task correctly. How-
ever, specic EMG patterns associated with the different phantom movements were not distinguishable because of the MRI-artefact contamination.
Thirty-two 5 mm thick axial slices were obtained with a T2* weighted gradient echo, echo planar imaging sequence, using blood oxygen level-
dependent contrast (repetition time 3030 ms, echo time 40 ms, ip angle of 90, matrix 64 64, eld of view 220 220 mm2), and 218 brain volumes
were acquired for each of the ve runs. The rst ve and the last three volumes of each run were discarded to ensure signal equilibrium.

Image processing

Image processing was carried out using Statistical Parametric Mapping (SPM 8, Wellcome Department of Cognitive Neurology, London, UK, www.
l.ion.ucl.ac.uk/spm), implemented in MATLAB 7.6 (Mathworks, Natick, MA, USA). Preprocessing included 3D motion correction and spatial smooth-
ing with a Gaussian kernel 8 8 8 mm3 full width at half maximum. Functional datasets were then co-registered with the T1-weighted anatomical
scan and transformed into standard MNI space using a segmentation-based normalization approach (Ashburner and Friston, 2005). The realignment
parameters estimated during spatial pre-processing were introduced in the design matrix as regressors of no interest in order to prevent confounding
activations related to minor head movements during scanning (Samii et al., 1996). Volumes in which head motion exceeded a threshold were with-
held from GLM estimation. Motion censoring was performed with a framewise displacement threshold of 0.90 mm (Siegel et al., 2014) and volumes
in which head motion exceeded 0.9 mm were withheld from GLM estimation. This threshold was chosen to remove modest portions of motion-
contaminated data, not to remove all volumes with motion. The experimental block design was convolved with the canonical hemodynamic re-
sponse function, and the resulting model was estimated using a highpass lter at 128 s to remove low-frequency artefacts.

Individual and group activation maps

In a rst-level analysis, contrast maps were calculated for the active periods versus rest separately for each run and for each subject. Analyses were
performed with a threshold of p b 0.05, corrected for multiple comparisons using Family Wise Error (FWE). Statistical results were mapped onto each
subject's 3D mesh of the cortical surface. Cortical surface reconstruction was generated from T1-weighted images using the standard procedure pro-
vided by Caret 5.562 software (http://brainvis.wustl.edu) (Van Essen et al., 2001). A single Region Of Interest (ROI) was then manually dened on
 E. Rafn et al. / NeuroImage 130 (2016) 134144 137

each subject's attened anatomic image by a trained investigator (E.R.) and checked by the co-authors (P.G. and N.R.). The ROI included the complete
anterior wall of the central sulcus which includes BA 4p and the posterior part of BA 4a (Geyer et al., 1996).
Since not all amputations were on the same side of the body, data from the ve amputees with a right-side amputation were ipped (left-to-right)
so that all amputees were considered as left hand amputated and right hand intact (this method has been used for example by Diers et al., 2010;
MacIver et al., 2008). In order to ensure that this side-standardization procedure was appropriate for our data set we performed a random effect anal-
ysis (two sample t-test) in which we compared ipped right hand amputees (n = 3) with the non-ipped left hand amputees (n = 8). This was done
for both elbow and hand movements and for both sides of the body. There were no differences in either hemisphere between the two groups for any
of the comparisons, and all subsequent analyses of the amputee group were performed with all amputees considered as left hand amputated.
Before pooling the data from the above- and below-elbow amputees we performed a random effect with a t-test analysis on activation related to
amputated side hand and elbow movements, comparing above-elbow (n = 8) and below-elbow (n = 3) amputees. This analysis was motivated by
the fact that phantom hand movements in above-elbow amputees are associated with contractions in muscles situated proximal to the elbow (e.g.
Biceps, Triceps or Deltoid muscles) whereas in below-elbow amputees they involve contractions in extrinsic hand muscles situated in the forearm.
Regardless of amputation level all phantom elbow movements involved contractions in the biceps and triceps, but in below-elbow amputees these
muscles were intact and the movement involved actual movement of the elbow, whereas in above-elbow amputees the muscles were partially am-
putated and the movement was a phantom movement as the elbow was not present. This analysis revealed no differences when we contrasted hand
or elbow movements in the two sub-groups, which prompted us to proceed, as is commonly done in the literature (see Bogdanov et al., 2012; Diers
et al., 2010; Grusser et al., 2001; Makin et al., 2015a), with a single group analysis combining both above and below elbow amputees (two sample t-
tests p b 0.05 corrected for multiple comparisons, FWE).
In a second-level analysis, we performed a random effect analysis (two sample t-test) with the side-standardized individual activation maps using
the baseline contrasts of task versus rest. Results were thresholded at a p value less than 0.05 corrected for multiple comparisons using FWE.

Intensity, volume and overlap of M1-c clusters

The Marsbar-toolbox was used to extract the percent signal change relative to the mean signal from each individual M1-c ROIs (http://marsbar.
sourceforge.net) (Brett et al., 2002).
For each subject we also quantied the volume of activation in M1-c generated by each of the ve movements by counting the number of
activated voxels in each M1-c after thresholding the statistical map. We also estimated the degree of overlap between the activations associated
with any two given movements (Mvt 1 and Mvt 2 in the following formula) using a custom-built program implemented in MATLAB:

No: of common voxels  100

Percent of overlap
Total no: of voxels Mvt 1 Total no: of voxels Mvt 22  No: of common voxels

Activation volume and percent of overlap were computed from the individual maps (thresholded at p b 0.05 and corrected for multiple compar-
isons (FWE)).

Activation shifts in M1-c

Since a lot of studies examining reorganization after amputation report topological asymmetries in the sensorimotor cortices, we also examined
asymmetries in the Centre of Gravity (CoG) of lip, elbow, and hand movement activations. We identied the M1-c locus associated with each move-
ment by rst computing the CoG of each movement's activation on the cortical surface. The geodesic distance between each CoG and the most medial
point of M1-c was then calculated on each individual's cortical mesh using the Dijktra algorithm (Dijkstra, 1959), which computes the shortest path
between two points (see supplementary methods for a complete description of our cortical-surface-based method), and this distance was expressed

Fig. 1. CoGs for the elbow, hand and lip for one control subject in the left and the right M1-c. Most of the typical tools used to characterize cortical organization are based on a 3-dimensional
referential and Euclidian distances. This gure highlights the convoluted surface of the primary motor cortex and the longer route of the geodesic distance used in our method as compared
to a straight Euclidian pathway. This surface-based assessment gives a more realistic topographic localization of activity within the primary motor cortex.
138 E. Rafn et al. / NeuroImage 130 (2016) 134144

as a percentage of the overall length of the subject's M1-c in the medio-lateral axis (0% = CoG positioned in the interhemispheric sulcus; 100% = CoG
positioned at the lateral extremity of M1-c).
Fig. 1 shows the CoG positions of the elbow, hand, and lip movements in the left and right motor cortices of one control subject. We compared the
positions of the CoGs in the intact and amputated hemispheres of amputees. We also evaluated whether there were any differences in absolute CoG
position between controls and M1-cPHANTOM or M1-cINTACT. Note that all these measures (morphometric and functional) were computed in a more re-
alistic way than in many previous studies because they were evaluated on the reconstructed cortical surface (2D) from 3D MRI (Anticevic et al., 2008;
Oosterhof et al., 2011).

Statistical analyses
Between-group comparisons were performed between M1-cPHANTOM and M1-cNON-DOM and between M1-cINTACTAND M1-cDOM (see e.g. Hahamy et al.,
2015; Makin et al., 2015a, 2015b). All statistical analyses were conducted with a signicance level set at 0.05 using STATISTICA 8.0 (StatSoft, Inc.). The
KolmogorovSmirnov test showed that not all our variables followed a normal distribution (p N 0.05). Given this, and our sample size (Conover,
1999), we chose to assess within-group differences using the Wilcoxon rank sum test and between-group differences using the MannWhitney U
test. Throughout the manuscript medians are reported with the range indicated in brackets. The hemisphere contralateral to the amputation
(M1-cPHANTOM) was compared to the non-dominant hemisphere of the controls (M1-cNon-Dom.) and the hemisphere contralateral to the intact limb
(M1-cINTACT), to the dominant hemisphere (M1-cDOM.) (see Makin et al., 2013a). Finally, we used a bootstrapping procedure to verify that our results
held true when the control and amputee groups were the same size (n = 11) (see supplementary materials and Table S1 for results).

Results
Clinical characteristics of amputees
Our sample of amputees was heterogeneous in terms of chronic
Phantom Limb Pain intensity: three had a high level of pain (6 out of
10 on the VAS), six had moderate pain, and two had no phantom limb
pain at all. Patients also varied in the time since their amputation, but
all reported stable pain levels over the three months prior to testing,
suggesting that their pain was chronic. The average constant pain inten-
sity for the group was 4/10. The average speed when performing the
nger-to-thumb opposition task was 0.18 cycles per second [0.09
0.34]. Clinical data for all amputees are reported in Table 1.
M1 morphometry
We compared the surface area (mm2), depth (mm) and thickness
(mm) of the central sulcus between both hemispheres of amputees and
controls (Fig. 2). The amputated and intact hemispheres did not differ
on any of the three measures. In control subjects the Central Sulcus was
deeper in the dominant hemisphere, consistent with previous studies
(Cykowski et al., 2008; Davatzikos and Bryan, 2002; Manceaux-Demiau
et al., 1998; Sun et al., 2011). The amputated hemisphere was thinner
and smaller than the non-dominant hemisphere of controls. Medians
and interquartile ranges as well as statistical data are shown in Fig. 2.
Individual and group activation maps
The individual activation maps (rst-level analysis, not shown) re-
vealed that the contralateral primary sensory and motor cortices (S1-c
and M1-c) were signicantly activated in all 11 amputees and in all 17
control subjects for the ve different movements. Although activity
was less consistent across subjects, signicant clusters were present in
the ipsilateral M1 and in non-primary sensorimotor areas including
the bilateral Supplementary Motor Area (SMA), the bilateral dorsal
premotor cortex (dPMC), the cingulate motor area, the bilateral inferior
and superior parietal cortices, and the ipsilateral cerebellum.
The results of the random effect group analysis in the hemisphere
contralateral to each movement (except for lip movements for which
activations are represented bilaterally) are displayed in Fig. 3 (control
subjects on the left and amputees on the right). For both groups, activity
was present in S1-c, M1-c, SMA, and dPMC during all movements. With-
in M1-c each movement's representation was roughly positioned in the
classical somatotopic order (Peneld, 1937), with the elbow, the hand,
and the lips positioned progressively more laterally (see Table 2 and
see for example Alkadhi et al., 2002; Grafton et al., 1993; Lotze et al.,
2000; Plow et al., 2010 for condence intervals published on
somatotopic motor activations).
BOLD responses during hand movements
Hand: activation intensity, cluster volume, and CoG position
Activation volumes in M1-cPHANTOM were similar to M1-cNON-DOM but
smaller than in M1-cINTACT, meaning that activation volume in the contra-
lateral M1 was very high when moving the intact hand (Fig. 4A). CoG
position in M1-cPHANTOM did not differ from CoG position in M1-cNON-DOM
or M1-cINTACT, and CoG position in M1-cINTACT and M1-cDOM did not differ
from each other (Fig. 4B). The percent signal change did not differ be-
tween M1-cPHANTOM and M1-cINTACT (Z = 0.09, p = 0.92), M1-cPHANTOM and
M1-cNON-DOM (U = 61, p = 0.13), or M1-cINTACT and M1-cDOM (U = 78,

Fig. 2. Medians and interquartile ranges of the CS surface (mm2) (A), CS depth (mm) (B) and CS thickness (mm) (C) in both hemispheres of amputees (M1-cPHANTOM/M1-cINTACT) and both
hemispheres of controls (M1-cDOM/M1-cNON-DOM); CS surface: M1-cPHANTOM versus M1-cINTACT: Z = 1.13, p = 0.63, M1-cPHANTOM versus M1-cNON-DOM: U = 43, p = 0.04, M1-cINTACT versus M1-cDOM:
U = 83, p = 0.64; CS depth: M1-cPHANTOM versus M1-cINTACT: Z = 1.83, p = 0.07, M1-cPHANTOM versus M1-cNON-DOM: U = 65, p = 0.25, M1-cINTACT versus M1-cDOM: U = 71, p = 0.42, M1-cDOM versus
M1-cNON-DOM: Z = 2.67, p = 0.01; CS thickness: M1-cPHANTOM versus M1-cINTACT: Z = 0, p = 1, M1-cPHANTOM versus M1-cNON-DOM: U = 0, p b 0.001, M1-cINTACT versus M1-cDOM: U = 48.5, p = 0.08,
M1-cDOM versus M1-cNON-DOM: Z = 0.92, p = 0.36.
 E. Rafn et al. / NeuroImage 130 (2016) 134144 139

Fig. 3. Random effect results in the hemisphere contralateral to the movements (except for lip movements which are represented bilaterally). Elbow, hand, and lip activations in M1-c
(pink zone) for the right and left limbs of control subjects (left panel) and for the intact and phantom limbs of amputees (right panel) (p b 0.05, FWE corrected for multiple
comparisons). In both groups, activity was also present during all three movements in S1-c, SMA, and dPMC. This was(veried using MNI coordinates of activation peaks using the
automated anatomical labelling (AAL; Tzourio-Mazoyer et al., 2002) and Wake Forest University PickAtlas (WFU; Maldjian et al., 2003, 2004) SPM toolboxes. Ante.: anterior, Post.:
posterior.

p = 0.48). There were no inter-hemispheric differences in any of the
three measures for the controls: (activation volumes: Z = 1.4, p =
0.16; CoG position: Z = 1.49, p = 0.14; percent signal change: Z =
1.87, p = 0.06).
Even though we found no group-level difference between hand acti-
vation in M1-cNON-DOM and M1-cPHANTOM the persistent representation
model predicts a relationship between M1-cPHANTOM activity during
hand movements and phantom pain. To investigate this we correlated
phantom limb pain as well as the phantom limb motor performance
with activation volume and activation intensity, but found no relation-
ship between these variables (phantom limb pain and activation vol-
ume: r = 0.18, p = 0.6, intensity r = 0.4, p = 0.22; motor
abilities and activation volume: r = 0.43, p = 0.2, intensity: r = 0.45,
p = 0.17, Bonferroni-Sidak corrected).
A second prediction of the persistent representation model is that the
intact hand activates the former hand area (Makin et al., 2013a). Thus, we
also examined the ipsilateral motor cortex activity generated during
hand movements. We found that 73% of amputees recruited the ipsilater-
al motor cortex (M1-cINTACT) when moving their phantom hand and 55%
when moving their intact hand (M1-cPHANTOM) compared with only
18% in the control group. Among those subjects who activated their ipsi-
lateral motor cortex, Fig. 4C shows that the ipsilateral M1 activation clus-
ters in amputees were larger than in controls when amputees moved
either their intact or their amputated hand. Of interest, the amount of ip-
silateral activity in the former hand area while moving the intact hand
was strongly correlated with the ability to move the phantom (r =
0.87, p = 0.01, Bonferroni-Sidak corrected). Unlike activation volumes,
activation intensities did not differ between the three groups (Fig. 4D).

Table 2
Random effect results for the control and amputee groups: MNI coordinates of the Tmax, maximal Z scores, and number of activated voxels (kE) related to the elbow, hand, and lip acti-
vations in M1-c; Hem.: Hemisphere; L/R: Left/Right. Note that for amputees, data were ipped so that all images of the right hemisphere correspond to the hemisphere contralateral to the
amputated side and all images of the left hemisphere correspond to the hemisphere contralateral to the intact side.

Controls Amputees

Movement Hem. Tmax (x, y, z) Z max kE Movement Hem. Tmax (x, y, z) Z max kE

Right elbow L 30/27/63 7.98 40 Intact right elbow L 33/27/60 13.29 26

Left elbow R 33/27/63 8.90 21 Amputated left elbow R 36/27/60 15.07 43
Right hand L 39/21/63 11.51 44 Intact right hand L 39/21/60 9.03 52
Left hand R 42/18/60 10.79 47 Amputated left hand R 36/12/63 10.13 25
Lips L 54/12/42 8.44 11 Lips L 54/12/39 7.21 17
Lips R 57/6/42 5.16 7 Lips R 48/12/41 6.39 27

Hem., Hemisphere.
140 E. Rafn et al. / NeuroImage 130 (2016) 134144

Fig. 4. Medians and interquartile ranges of volumes of activation (No. of voxels) (A) and CoG positions (%) (B) in the contralateral motor cortex during phantom hand movements (M1-
cPHANTOM), intact hand movements (M1-cINTACT) and dominant (M1-cDOM) and non-dominant hands movements in controls (M1-cNON-DOM). Activation volumes: M1-cPHANTOM vs M1-cINTACT: Z =
2 p = 0.04, M1-cPHANTOM vs M1-cNON-DOM: U = 89, 0 = 0.85, M1-cINTACT vs M1-cDOM: U = 47, p = 0.03 and M1-cDOM versus M1-cNON-DOM: Z = 1.4, p = 0.16. CoG positions: M1-cPHANTOM vs M1-
cINTACT: Z = 0.07, p = 0.93, M1-cPHANTOM vs M1-cNON-DOM: U = 89, p = 0.85, M1-cINTACT vs M1-cDOM: U = 59, p = 0.11 and M1-cDOM versus M1-cNON-DOM: Z = 1.4, p = 0.16. Medians and
interquartile ranges of volumes of activation (No. of voxels) (C) and intensity of activation (% signal change) (D) in the ipsilateral motor cortex during phantom hand movements
(volume and intensity of activation in M1-cINTACT: kE = 39 [SEM = 8], percent signal change: 0.59 [SEM = 0.07]), during intact hand movements (volume and intensity of activation in
M1-cPHANTOM: kE = 25 [SEM = 10], percent signal change = 0.67 [SEM = 0.11]), and during dominant and non-dominant hands movements in controls (volumes and intensities of
activation, in M1-cNON-DOM: kE = 1 [SEM = 1], percent signal change: 0.77 [SEM = 0.1]; in M1-cDOM,: kE = 4 [SEM = 0.5], percent signal change = 0.47 [0.05]).

Hand: overlap with elbow and lip activations in M1-c
The average number of voxels activated during both hand and elbow
movements was greater in M1-cPHANTOM than in M1-cINTACT and M1-cNON-
DOM. There were also more overlapping voxels in M1-cINTACT than in
M1-cDOM (Fig. 5A). Very few voxels overlapped when hand and lip
activations were compared (range from 0 to 2.4%), but on average
more voxels were common to hand/lip movements in M1-cPHANTOM and
M1-cINTACT than in M1-cNON-DOM, but this difference was only signicant
between M1-cPHANTOM and M1-cNON-DOM (Fig. 5B).
BOLD responses during lip and elbow movements in M1-c
While the hand representation did not show any clear asymmetries
between the two hemispheres of the amputees nor between the ampu-
tees and the controls, the overlap analysis presented above suggests
that adjacent body parts representations underwent substantial reorga-
nization. This is clearly illustrated in Fig. 6A which shows that activation
volume during lip movements was signicantly larger in M1-cPHANTOM
than in M1-cINTACT and M1-cNON-DOM, whereas M1-cINTACT and M1-cDOM vol-
umes did not differ. It also shows a similar pattern for the CoG positions
(Fig. 6B), with the lip CoG situated more medially in M1-cPHANTOM than in
M1-cINTACT or M1-cNON-DOM, and in a similar mediolateral position in M1-
cINTACT and M1-cDOM. The reorganizational changes observed for the lips
were also present for the elbow: activation volume during elbow move-
ments was signicantly larger in M1-cPHANTOM than in M1-cINTACT and M1-
cNON-DOM, whereas M1-cINTACT and M1-cDOM volumes did not differ (Fig. 6C).
Similarly, the elbow CoG was situated more laterally in M1-cPHANTOM than
in M1-cINTACT or M1-cNON-DOM, and in a similar mediolateral position in M1-
cINTACT and M1-cDOM (Fig. 6D).
The percent signal change measure revealed no clear differences for
either the lip or elbow movements. Percent signal change during lip
movements was similar in M1-cPHANTOM, M1-cINTACT, and M1-cNON-DOM
(M1-cPHANTOM versus M1-cINTACT: Z = 1.38, p = 0.17, M1-cPHANTOM versus
M1-cNON-DOM: U = 60, p = 0.22) as for M1-cINTACT and M1-cDOM (U = 25,
p = 0.06). Similarly, percent signal change was similar during elbow

Fig. 5. Medians and interquartile ranges of overlap (%) between hand/elbow (A) and hand/lips (B) in both hemispheres of amputees (M1-cPHANTOM and M1-cINTACT) of controls (M1-cDOM and
M1-cNON-DOM); hand/elbow overlap: M1-cPHANTOM versus M1-cINTACT: Z = 2, p = 0.04, M1-cPHANTOM versus M1-cNON-DOM: U 0, p b 0.001, M1-cINTACT versus M1-cDOM: U = 18, p b 0.001, M1-cDOM
versus M1-cNON-DOM: Z = 0.59, p = 0.55; hand/lips overlap: M1-cPHANTOM versus M1-cINTACT: Z = 0.27, p = 0.79, M1-cPHANTOM versus M1-cNON-DOM: U = 42, p = 0.01, M1-cINTACT versus M1-cDOM:
U = 68, p = 0.33; M1-cDOM versus M1-cNON-DOM: Z = 1.02, p = 0.31.
 E. Rafn et al. / NeuroImage 130 (2016) 134144 141

Fig. 6. Medians and interquartile range of the activation volumes (no. of voxels) and CoGs positions (%) associated with lip movements (respectively A and B) and with elbows movements
(respectively C and D) in the hemisphere contralateral to the amputation (M1-cPHANTOM), the intact limb (M1-cINTACT) and both hemispheres of controls (M1-cDOM and M1-cNON-DOM). Lip volumes:
M1-cPHANTOM versus M1-cINTACT (Z = 2.93, p = 0.003), M1-cPHANTOM versus M1-cNON-DOM (U = 46, p = 0.04), M1-cINTACT versus M1-cDOM (U = 54, p = 0.1, M1-cDOM versus M1-cNON-DOM: Z = 1.42, p =
0.16). CoGs positions: M1-cPHANTOM versus M1-cINTACT (U = 2.93, p = 0.003), M1-cPHANTOM versus M1-cNON-DOM (U = 46, p = 0.04), M1-cINTACT versus M1-cCONTROL (U = 1.2, p = 0.23, M1-cDOM versus
M1-cNON-DOM: Z = 1.82, p = 0.07). Elbow volumes: M1-cPHANTOM versus M1-cINTACT (Z = 2.93, p = 0.003), M1-cPHANTOM versus M1-cNON-DOM (U = 35, p = 0.006), M1-cINTACT versus M1-cDOM (U = 71,
p = 0.3), M1-cDOM versus M1-cNON-DOM: Z = 0.72, p = 0.49. CoGs positions: M1-cPHANTOM versus M1-cINTACT (Z = 2.85, p = 0.004), M1-cPHANTOM versus M1-cNON-DOM (U = 43, p = 0.02), M1-cINTACT versus
M1-cDOM (U = 76, p = 0.42), M1-cDOM versus M1-cNON-DOM: Z = 0.5, p = 0.62.

movements in M1-cPHANTOM, and M1-cINTACT (Z = 0.44, p = 0.66) between Discussion

M1-cPHANTOM and M1-cNON-DOM (U = 71, p = 0.3) and between, M1-cINTACT
and M1-cDOM (U = 87, p = 0.77). We found clear asymmetries between the two hemispheres of
Lip and elbow movements had asymmetric representations in the upper-limb amputees for both lip and elbow representations, but not
amputated and intact hemispheres both in terms of activation volumes for the hand representation. Furthermore, both lip and elbow (but not
and medio-lateral CoG positions. While lip and elbow asymmetry did hand) representations differed between M1-cPHANTOM and the control
not correlate for either of these measures, lip and elbow CoG group, while hand (but not lip or elbow) representations differed be-
asymmetries both correlated signicantly with chronic phantom pain tween M1-cINTACT and the control group. Lip and elbow reorganization
and phantom limb motor performance, which themselves were signi- were not correlated with each other but were correlated with phantom
cantly correlated (r = 0.72, p = 0.012, Bonferroni-Sidak corrected) pain and phantom motor performance (with more reorganization asso-
(See Table 3). ciated with higher pain levels and poorer phantom motor control).
Thus, we suggest that the reorganization of these body parts might be
driven by common factors but that these factors do not equally reorga-
nize both body part representations. In addition to providing further
support for the maladaptative plasticity model our data also show that
Table 3
pain is not the only perceptual correlate of post-amputation reorganiza-
Spearman correlation coefcients (r) between clinical features and imaging data for lip tion, and that this reorganization is not limited to the face but extends to
and elbow movements in amputees. the medial part of the sensorimotor cortex.
Chronic phantom limb pain Phantom motor performance
(010) (cycles/s.) Hand, lip and elbow representations
Lip
Volume 0.21 0.11 The maintenance of an area within M1 devoted to the missing hand
Asymmetry (0.54) (0.75) is now well documented (Ersland et al., 1996; Makin et al., 2013b;
Mercier et al., 2006; Roux et al., 2001, 2003). In line with this, we ob-
Lip
CoG 0.74* 0.79* served that opening and closing the phantom hand produced substan-
Asymmetry (0.009) (0.004) tial activity in the primary motor cortex contralateral to the
amputation. Interestingly, not only was this activity symmetric with
Elbow
Volume 0.06 0.02 that recorded during intact hand movements, but it was also positioned
Asymmetry (0.87) (0.96) in the same relative location and had a similar volume and intensity as
activity during hand movements in control subjects.
Elbow
CoG 0.64* 0.63* Contrary to recent evidence suggesting a loss of grey matter in the
Asymmetry (0.035) (0.036) hand sensorimotor region contralateral to an amputation (Makin
CoG, Centre of Gravity.
et al., 2013b; Preissler et al., 2012; Xie et al., 2013), we observed no mor-
Asterisks represent signicant correlations and values in brackets represent uncorrected phological or functional asymmetries between the sensorimotor corti-
p-values. P values are provided for signicant correlations only. ces of our amputees. Possible reasons for this could be that our
142 E. Rafn et al. / NeuroImage 130 (2016) 134144

patients had relatively well-preserved phantom motor control and/or somatosensory and motor cortices, as has been demonstrated in a mon-
because they disproportionately recruited the deprived cortex when key (Schieber and Deuel, 1997) and human amputee (Ojemann and
moving their intact hand (Bogdanov et al., 2012; Cruz et al., 2003; Silbergeld, 1995).
Dettmers et al., 2001; Hamzei et al., 2001; MacIver et al., 2008; Makin
et al., 2013a). In accordance with limb-use patterns of traumatic ampu-
Phantom limb pain, phantom motor control and post-amputation
tees reported by Makin et al. (2013a, 2013b), most of our amputees re-
reorganization
lied exclusively on their intact hand for all manual tasks. Thus, this
activity pattern could be due to compensatory overuse of the intact
Although it has been suggested that the maladaptative plasticity
hand. Unfortunately, this idea remains speculative since we did not
model and the persistent representation model could co-exist in the pri-
measure compensatory every-day usage of the residual arm, as our pri-
mary somatosensory cortex (see the recent modelling study by Bostrom
mary goal was to depict reorganizational phenomena in relation to re-
et al., 2014), our data are not consistent with the persistent representa-
sidual phantom limb motor control and phantom limb levels. In light
tion model. Instead, we replicated the positive correlation between the
of recent work showing that reorganization patterns are associated
degree of lip reorganization and chronic phantom limb pain (Karl et al.,
with limb usage patterns (Makin et al., 2013a, 2013b, Hahamy et al.,
2001; Karl et al., 2004; Lotze et al., 2001; Lotze et al., 1999; MacIver et al.,
2015) future studies should monitor physical activity over a number
2008), and further demonstrated that reorganization of the medial por-
of days using objective measures like limb acceleration recordings. An
tion of M1 is also correlated with phantom limb pain. The bi-directional
alternative explanation for the high level of contralateral activity during
invasion of the former hand territory and its association with phantom
intact hand movements is that it arises from changes in inter-
limb pain extends the maladaptative plasticity model to include chang-
hemispheric inhibition. Since we observed a bilateral increase in activa-
es in both the medial and lateral portions of the cortex, and is coherent
tion when amputees moved their intact hand the unilateral deafferenta-
with other models of neuropathic pain. For example, the cortical repre-
tion combined with long-term increased use of the intact hand might
sentations of both the digits and the lips appear to shift and shrink in
have reduced inter-hemispheric inhibition (see also Makin et al.,
unilateral upper limb CRPS1 and the extent of their shift is correlated
2013b). This disinhibition hypothesis is congruent with decreased
with pain intensity (Maihofner et al., 2003, 2004).
callosal white-matter fractional anisotropy as measured with
While we cannot dissociate the relative contributions of phantom
Diffusion-weighted MR imaging (Simoes et al., 2012).
limb pain and phantom motor control to post-amputation reorganiza-
Activity associated with lip and elbow movements was about 30%
tion, several reasons suggest that phantom motor control plays a role
larger and was shifted by 5% along the central sulcus (medially for the
in reorganizational changes. First, data from patients with movements
lips and laterally for the elbow) in the hemisphere contralateral to the
disorders such as writer's cramp, motor stroke, cerebral palsy, or
amputation. The expansion of the lips' motor representation toward
Parkinson's disease show that disorganized upper limb representations
the deafferented hand region has been demonstrated in animals
are associated with decreased distal motor control (Kesar et al., 2012;
(Huntley, 1997; Kambi et al., 2011), in a number of PET and fMRI studies
Lindenbach and Bishop, 2013; Tyc et al., 2012; Yao et al., 2009). Second,
in humans (e.g. Karl et al., 2001; Karl et al., 2004; Kew et al., 1994; Lotze
there is increasing evidence for the efcacy of non-pharmacological
et al., 2001) and in some TMS studies (Dettmers et al., 1999; Roricht
antalgic therapies such as phantom limb training or mirror therapy
et al., 2001). The invasion of both face and stump representations into
(Chan et al., 2007; MacIver et al., 2008; Mercier and Sirigu, 2009). In-
the deprived cortex has been reported previously in animals (Wu and
deed, one of the rst studies on the use of mirror therapy reported im-
Kaas, 1999).
proved phantom movement capacities and pain relief after mirror
Motor cortex activity triggered by voluntary movements necessarily
therapy (Ramachandran and Rogers-Ramachandran, 1996), and recent
encompasses some sensory activity generated by the afferent signals
studies have demonstrated that mental imagery or mirror therapy can
produced by the movements themselves. Since the topological organi-
decrease pain and restore motor cortex organization (Foell et al.,
sation of the sensory cortex is altered after amputation all studies that
2014; Giraux and Sirigu, 2003; MacIver et al., 2008; Rothgangel et al.,
use voluntary movements to examine reorganisation of the motor cor-
2011).
tex inevitably measure changes within the motor cortex itself as well
Finally, our nding of substantial activity in the deprived motor cor-
as the reection of the sensory cortex reorganisation. It is important to
tex during intact hand movements, which is strongly correlated with
note, however, that mapping techniques unaffected by sensory re-
the ability to move the phantom, suggests that the synergistic move-
afference signals have shown changes within the motor cortex indepen-
ment of the intact and phantom limbs during mirror or imagery thera-
dent of sensory cortex changes (e.g. Kew et al., 1994; Schwenkreis et al.,
pies might help shape and reactivate the deprived cortex. This would
2001; Karl et al., 2001; Ojemann and Silbergeld, 1995). Our study is the
in turn facilitate voluntary phantom hand movements, protect against
rst to demonstrate reorganization of both the lip and elbow within the
reorganization of the motor cortex contralateral to the amputation,
motor cortex in a group of human upper-limb amputees (see Giraux
and thereby reduce phantom pain.
et al., 2001 for a single case study), and we suggest that these changes
reect reorganisation within the motor cortex itself as well as a possible
contribution from sensory cortex reorganisation. Conclusions
Our nding of a smaller hand/lips overlap compared to the hand/
elbow overlap and the absence of full invasion (despite the medial Our investigation of the functional and structural organization of
shift) of the lips into the hand territory reported by Makin et al. motor cortex in upper-limb amputees suggests that post-amputation
(2015a, 2015b) are consistent with animal data. Qi et al. (2000) report- reorganization is a complex phenomenon and that reorganization of dif-
ed that only a small number of stimulation sites in the former hand area ferent body parts might be driven by a combination of various underly-
of four monkeys evoked face movements. They reasoned that the rela- ing factors giving rise to sensorimotor phantom limb sensations. As poor
tively clear boundary between the face and hand representations motor control over the phantom limb was related to more reorganiza-
(Fang et al., 2002; Samulack et al., 1990) restricted cross-border plastic- tion, further research is needed in order to disentangle the relative con-
ity between the hand and the lips. However, this is difcult to reconcile tributions of phantom motor control and phantom limb pain to post-
with the nding, in the sensory domain, of the presence of referred sen- amputation reorganization. In addition to furthering our understanding
sations between the phantom hand and the face (Ramachandran et al., of the mechanisms underlying mirror therapy, this knowledge could
1992) and with the recent nding that touch improvement at the n- also contribute to our understanding of the therapeutic effects of other
gertip transfers to the lips (Muret et al., 2014). This might be due to dif- forms of therapies like repetitive transcranial magnetic stimulation of
ferences in the reorganizational processes occurring in the the motor cortex (Di Rollo and Pallanti, 2011; Li et al., 2009; Topper
 E. Rafn et al. / NeuroImage 130 (2016) 134144
et al., 2003) or direct electrical stimulation of the precentral gyrus
(Kleiser et al., 2010; Roux et al., 2008; Saitoh et al., 1999).
Funding
This work was supported by the Labex/Idex ANR-11-LABX-0042 and
IHU CeSaMe ANR-10-IBHU-0003 and ER was nanced by the French
Ministry of Health (PHRC 2007 N_0701040) awarded to PG, and by a
Interdisciplinary Synergy Program sponsored by the Novonordisk
Foundation (Biophysically adjusted state-informed cortex stimulation
(BASICS)).
Acknowledgments
We would like to thank D. Ibarrola, J.-C. Comte and D. Sappey-
Mariner at the CERMEP-Imagerie du vivant research imaging centre
F-69675 Bron, France for providing MRI facilities and for their technical
support. We also thank Dr. Angela Sirigu, Centre de Neurosciences
Cognitives, CNRS, 67, Blv. Pinel, F-69675 Bron, France for nancing the
control subject fMRI scans.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.neuroimage.2016.01.063.
References
Alkadhi, H., Crelier, G.R., Boendermaker, S.H., Golay, X., Hepp-Reymond, M.C., Kollias, S.S.,
2002. Reproducibility of primary motor cortex somatotopy under controlled condi-
tions. AJNR Am. J. Neuroradiol. 23, 15241532.
Anticevic, A., Dierker, D.L., Gillespie, S.K., Repovs, G., Csernansky, J.G., Van Essen, D.C.,
Barch, D.M., 2008. Comparing surface-based and volume-based analyses of functional
neuroimaging data in patients with schizophrenia. NeuroImage 41, 835848.
Ashburner, J., Friston, K.J., 2005. Unied segmentation. NeuroImage 26, 839851.
Birbaumer, N., Lutzenberger, W., Montoya, P., Larbig, W., Unertl, K., Topfner, S., Grodd, W.,
Taub, E., Flor, H., 1997. Effects of regional anesthesia on phantom limb pain are
mirrored in changes in cortical reorganization. J Neurosci 17, 55035508.
Bogdanov, S., Smith, J., Frey, S.H., 2012. Former hand territory activity increases after
amputation during intact hand movements, but is unaffected by illusory visual
feedback. Neurorehabil. Neural Repair.
Bostrom, K.J., de Lussanet, M.H., Weiss, T., Puta, C., Wagner, H., 2014. A computational
model unies apparently contradictory ndings concerning phantom pain. Sci. Rep.
4, 5298.
Brett, M., Anton, J.L., Valabregure, R., Poline, J.B., 2002. Region of interest analysis using
SPM toolbox. 8th International Conferance on Functional Mapping of the Human
Brain. Sendai, Japon: NeuroImage.
Brunelli, S., Morone, G., Iosa, M., Ciotti, C., De Giorgi, R., Foti, C., Traballesi, M., 2015.
Efcacy of progressive muscle relaxation, mental imagery, and phantom exercise
training on phantom limb: a randomized controlled trial. Arch. Phys. Med. Rehabil.
96, 181187.
Chan, B.L., Witt, R., Charrow, A.P., Magee, A., Howard, R., Pasquina, P.F., Heilman, K.M.,
Tsao, J.W., 2007. Mirror therapy for phantom limb pain. N. Engl. J. Med. 357,
22062207.
Chen, A., Yao, J., Kuiken, T., Dewald, J.P., 2013. Cortical motor activity and reorganization
following upper-limb amputation and subsequent targeted reinnervation.
Neuroimage Clin. 3, 498506.
Conover, W.J., 1999. Practical Nonparametric Statistics (Vol. Third Edition).
Cruz, V.T., Nunes, B., Reis, A.M., Pereira, J.R., 2003. Cortical remapping in amputees and
dysmelic patients: a functional MRI study. NeuroRehabilitation 18, 299305.
Cykowski, M.D., Coulon, O., Kochunov, P.V., Amunts, K., Lancaster, J.L., Laird, A.R., Glahn,
D.C., Fox, P.T., 2008. The central sulcus: an observer-independent characterization
of sulcal landmarks and depth asymmetry. Cereb. Cortex 18, 19992009.
Davatzikos, C., Bryan, R.N., 2002. Morphometric analysis of cortical sulci using parametric
ribbons: a study of the central sulcus. J. Comput. Assist. Tomogr. 26, 298307.
Dettmers, C., Liepert, J., Adler, T., Rzanny, R., Rijntjes, M., van Schayck, R., Kaiser, W.,
Bruckner, L., Weiller, C., 1999. Abnormal motor cortex organization contralateral to
early upper limb amputation in humans. Neurosci. Lett. 263, 4144.
Dettmers, C., Adler, T., Rzanny, R., van Schayck, R., Gaser, C., Weiss, T., Miltner, W.H.,
Bruckner, L., Weiller, C., 2001. Increased excitability in the primary motor cortex
and supplementary motor area in patients with phantom limb pain after upper
limb amputation. Neurosci. Lett. 307, 109112.
Di Rollo, A., Pallanti, S., 2011. Phantom limb pain: low frequency repetitive transcranial
magnetic stimulation in unaffected hemisphere. Case Rep. Med. 2011, 130751.
Diers, M., Christmann, C., Koeppe, C., Ruf, M., Flor, H., 2010. Mirrored, imagined and exe-
cuted movements differentially activate sensorimotor cortex in amputees with and
without phantom limb pain. Pain 149, 296304.
143
Dijkstra, E.W., 1959. A note on two problems in connexion with graphs. Numer. Math. 1,
269271.
Elbert, T., Sterr, A., Flor, H., Rockstroh, B., Knecht, S., Pantev, C., Wienbruch, C., Taub, E.,
1997. Input-increase and input-decrease types of cortical reorganization after upper
extremity amputation in humans. Exp. Brain Res. 117, 161164.
Ersland, L., Rosen, G., Lundervold, A., Smievoll, A.I., Tillung, T., Sundberg, H., Hugdahl, K.,
1996. Phantom limb imaginary ngertapping causes primary motor cortex
activation: an fMRI study. Neuroreport 8, 207210.
Fang, P.C., Jain, N., Kaas, J.H., 2002. Few intrinsic connections cross the hand-face border of
area 3b of new world monkeys. J. Comp. Neurol. 454, 310319.
Flor, H., 2002. The modication of cortical reorganization and chronic pain by sensory
feedback. Appl. Psychophysiol. Biofeedback 27, 215227.
Flor, H., Elbert, T., Knecht, S., Wienbruch, C., Pantev, C., Birbaumer, N., Larbig, W., Taub, E.,
1995. Phantom-limb pain as a perceptual correlate of cortical reorganization follow-
ing arm amputation. Nature 375, 482484.
Flor, H., Elbert, T., Muhlnickel, W., Pantev, C., Wienbruch, C., Taub, E., 1998. Cortical
reorganization and phantom phenomena in congenital and traumatic upper-
extremity amputees. Exp. Brain Res. 119, 205212.
Flor, H., Denke, C., Schaefer, M., Grusser, S., 2001. Effect of sensory discrimination training
on cortical reorganisation and phantom limb pain. Lancet 357, 17631764.
Foell, J., Bekrater-Bodmann, R., Diers, M., Flor, H., 2014. Mirror therapy for phantom limb
pain: brain changes and the role of body representation. Eur. J. Pain 18, 729739.
Gagn, M., Reilly, K.T., Hetu, S., Mercier, C., 2009. Motor control over the phantom limb in
above-elbow amputees and its relationship with phantom limb pain. Neuroscience
162, 7886.
Geyer, S., Ledberg, A., Schleicher, A., Kinomura, S., Schormann, T., Burgel, U., Klingberg, T.,
Larsson, J., Zilles, K., Roland, P.E., 1996. Two different areas within the primary motor
cortex of man. Nature 382, 805807.
Giraux, P., Sirigu, A., 2003. Illusory movements of the paralyzed limb restore motor cortex
activity. NeuroImage 20 (Suppl. 1), S107S111.
Giraux, P., Sirigu, A., Schneider, F., Dubernard, J.M., 2001. Cortical reorganization in motor
cortex after graft of both hands. Nat. Neurosci. 4, 691692.
Grafton, S.T., Woods, R.P., Mazziotta, J.C., 1993. Within-arm somatotopy in human motor
areas determined by positron emission tomography imaging of cerebral blood ow.
Exp. Brain Res. 95, 172176.
Grusser, S.M., Winter, C., Muhlnickel, W., Denke, C., Karl, A., Villringer, K., Flor, H., 2001.
The relationship of perceptual phenomena and cortical reorganization in upper
extremity amputees. Neuroscience 102, 263272.
Hahamy, A., Sotiropoulos, S.N., Henderson Slater, D., Malach, R., Johansen-Berg, H., Makin,
T.R., 2015. Normalisation of brain connectivity through compensatory behaviour,
despite congenital hand absence. Elife 4.
Hamzei, F., Liepert, J., Dettmers, C., Adler, T., Kiebel, S., Rijntjes, M., Weiller, C., 2001.
Structural and functional cortical abnormalities after upper limb amputation during
childhood. Neuroreport 12, 957962.
Huntley, G.W., 1997. Correlation between patterns of horizontal connectivity and the
extend of short-term representational plasticity in rat motor cortex. Cereb. Cortex
7, 143156.
Huse, E., Larbig, W., Flor, H., Birbaumer, N., 2001. The effect of opioids on phantom limb
pain and cortical reorganization. Pain 90, 4755.
Jiang, G., Yin, X., Li, C., L. L., Zhao, L., Evans, A.C., Jiang, T., Wu, J., Wang, J., 2015. The
plasticity of brain gray matter and white matter following lower limb amputation.
Neural Plast.
Kambi, N., Tandon, S., Mohammed, H., Lazar, L., Jain, N., 2011. Reorganization of the
primary motor cortex of adult macaque monkeys after sensory loss resulting from
partial spinal cord injuries. J. Neurosci. 31, 36963707.
Karl, A., Birbaumer, N., Lutzenberger, W., Cohen, L.G., Flor, H., 2001. Reorganization of
motor and somatosensory cortex in upper extremity amputees with phantom limb
pain. J. Neurosci. 21, 36093618.
Karl, A., Muhlnickel, W., Kurth, R., Flor, H., 2004. Neuroelectric source imaging of steady-
state movement-related cortical potentials in human upper extremity amputees with
and without phantom limb pain. Pain 110, 90102.
Kesar, T.M., Sawaki, L., Burdette, J.H., Cabrera, M.N., Kolaski, K., Smith, B.P., O'Shea, T.M.,
Koman, L.A., Wittenberg, G.F., 2012. Motor cortical functional geometry in cerebral
palsy and its relationship to disability. Clin. Neurophysiol. 123, 13831390.
Kew, J.J., Ridding, M.C., Rothwell, J.C., Passingham, R.E., Leigh, P.N., Sooriakumaran, S.,
Frackowiak, R.S., Brooks, D.J., 1994. Reorganization of cortical blood ow and trans-
cranial magnetic stimulation maps in human subjects after upper limb amputation.
J. Neurophysiol. 72, 25172524.
Kleiser, R., Staempi, P., Valavanis, A., Boesiger, P., Kollias, S., 2010. Impact of fMRI-guided
advanced DTI ber tracking techniques on their clinical applications in patients with
brain tumors. Neuroradiology 52, 3746.
Knecht, S., Henningsen, H., Elbert, T., Flor, H., Hohling, C., Pantev, C., Taub, E., 1996. Reorga-
nizational and perceptional changes after amputation. Brain 119 (Pt 4), 12131219.
Li, X., Yu, R.T., Xu, K., Li, F.C., Fan, Y.C., Gao, W.C., Guo, K.Q., Pan, X., Yang, C., 2009.
Application of diffusion tensor imaging in preoperation and postoperation patients
of glioma with 3.0 Tesla MRI. Zhonghua Yi Xue Za Zhi 89, 13001304.
Lindenbach, D., Bishop, C., 2013. Critical involvement of the motor cortex in the
pathophysiology and treatment of Parkinson's disease. Neurosci. Biobehav. Rev. 37,
27372750.
Lotze, M., Grodd, W., Birbaumer, N., Erb, M., Huse, E., Flor, H., 1999. Does use of a myoelec-
tric prosthesis prevent cortical reorganization and phantom limb pain? Nat. Neurosci.
2, 501502.
Lotze, M., Erb, M., Flor, H., Huelsmann, E., Godde, B., Grodd, W., 2000. fMRI evaluation of
somatotopic representation in human primary motor cortex. NeuroImage 11, 473481.
Lotze, M., Flor, H., Grodd, W., Larbig, W., Birbaumer, N., 2001. Phantom movements and
pain. An fMRI study in upper limb amputees. Brain 124, 22682277.
144 E. Rafn et al. / NeuroImage 130 (2016) 134144
MacIver, K., Lloyd, D.M., Kelly, S., Roberts, N., Nurmikko, T., 2008. Phantom limb pain, cor-
tical reorganization and the therapeutic effect of mental imagery. Brain 131,
21812191.
Maihofner, C., Handwerker, H.O., Neundorfer, B., Birklein, F., 2003. Patterns of cortical re-
organization in complex regional pain syndrome. Neurology 61, 17071715.
Maihofner, C., Handwerker, H.O., Neundorfer, B., Birklein, F., 2004. Cortical reorganization
during recovery from complex regional pain syndrome. Neurology 63, 693701.
Makin, T.R., Cramer, A.O., Scholz, J., Hahamy, A., Henderson Slater, D., Tracey, I., Johansen-
Berg, H., 2013a. Deprivation-related and use-dependent plasticity go hand in hand.
Elife 2, e01273.
Makin, T.R., Scholz, J., Filippini, N., Henderson Slater, D., Tracey, I., Johansen-Berg, H.,
2013b. Phantom pain is associated with preserved structure and function in the
former hand area. Nat. Commun. 4, 1570.
Makin, T.R., Filippini, N., Duff, E.P., Henderson Slater, D., Tracey, I., Johansen-Berg, H.,
2015a. Network-level reorganisation of functional connectivity following arm ampu-
tation. NeuroImage 114, 217225.
Makin, T.R., Scholz, J., Henderson Slater, D., Johansen-Berg, H., Tracey, I., 2015b.
Reassessing cortical reorganization in the primary sensorimotor cortex following
arm amputation. Brain.
Maldjian, J.A., Laurienti, P.J., Kraft, R.A., Burdette, J.H., 2003. An automated method for
neuroanatomic and cytoarchitectonic atlas-based interrogation of fMRI data sets.
NeuroImage 19, 12331239.
Maldjian, J.A., Laurienti, P.J., Burdette, J.H., 2004. Precentral gyrus discrepancy in electronic
versions of the Talairach atlas. NeuroImage 21, 450455.
Manceaux-Demiau, A., Bryan, R.N., Davatzikos, C., 1998. A probabilistic ribbon model for
shape analysis of the cerebral sulci: application to the central sulcus. J. Comput. Assist.
Tomogr. 22, 962971.
Mercier, C., Sirigu, A., 2009. Training with virtual visual feedback to alleviate phantom
limb pain. Neurorehabil. Neural Repair 23, 587594.
Mercier, C., Reilly, K.T., Vargas, C.D., Aballea, A., Sirigu, A., 2006. Mapping phantom move-
ment representations in the motor cortex of amputees. Brain 129, 22022210.
Montoya, P., Ritter, K., Huse, E., Larbig, W., Braun, C., Topfner, S., Lutzenberger, W., Grodd,
W., Flor, H., Birbaumer, N., 1998. The cortical somatotopic map and phantom phe-
nomena in subjects with congenital limb atrophy and traumatic amputees with
phantom limb pain. Eur. J. Neurosci. 10, 10951102.
Muret, D., Dinse, H.R., Macchione, S., Urquizar, C., Farne, A., Reilly, K.T., 2014. Touch im-
provement at the hand transfers to the face. Curr. Biol. 24, R736R737.
Ojemann, J.G., Silbergeld, D.L., 1995. Cortical stimulation mapping of phantom limb
rolandic cortex. Case report, J. Neurosurg. 82, 641644.
Oosterhof, N.N., Wiestler, T., Downing, P.E., Diedrichsen, J., 2011. A comparison of volume-
based and surface-based multi-voxel pattern analysis. NeuroImage 56, 593600.
Peneld, W.B.,.E., 1937. Somatic motor and sensory representation in the cerebral cortex
of man as studied by electrical stimulation. Brain 60, 389443.
Perrot, M., Riviere, D., Mangin, J.F., 2011. Cortical sulci recognition and spatial normaliza-
tion. Med. Image Anal. 15, 529550.
Plow, E.B., Arora, P., Pline, M.A., Binenstock, M.T., Carey, J.R., 2010. Within-limb
somatotopy in primary motor cortexrevealed using fMRI. Cortex 46, 310321.
Preissler, S., Feiler, J., Dietrich, C., Hofmann, G.O., Miltner, W.H., Weiss, T., 2012. Gray mat-
ter changes following limb amputation with high and low intensities of phantom
limb pain. Cereb. Cortex.
Qi, H.X., Stepniewska, I., Kaas, J.H., 2000. Reorganization of primary motor cortex in adult
macaque monkeys with long-standing amputations. J. Neurophysiol. 84, 21332147.
Rafn, E., Giraux, P., Reilly, K.T., 2011. The Moving Phantom: Motor Execution or Motor
Imagery? Cortex.
Rafn, E., Mattout, J., Reilly, K.T., Giraux, P., 2012a. Disentangling motor execution from
motor imagery with the phantom limb. Brain 135, 582595.
Rafn, E., Giraux, P., Reilly, K.T., 2012b. The moving phantom: motor execution or motor
imagery? Cortex 48, 746757.
Ramachandran, V.S., Rogers-Ramachandran, D., 1996. Synaesthesia in phantom limbs in-
duced with mirrors. Proc. Biol. Sci. 263, 377386.
Ramachandran, V.S., Rogers-Ramachandran, D., Stewart, M., 1992. Perceptual correlates of
massive cortical reorganization. Science 258, 11591160.
Reilly, K.T., Sirigu, A., 2008. The motor cortex and its role in phantom limb phenomena.
Neuroscientist 14, 195202.
Roricht, S., Machetanz, J., Irlbacher, K., Niehaus, L., Biemer, E., Meyer, B.U., 2001. Reorgani-
zation of human motor cortex after hand replantation. Ann. Neurol. 50, 240249.
Rothgangel, A.S., Braun, S.M., Beurskens, A.J., Seitz, R.J., Wade, D.T., 2011. The clinical as-
pects of mirror therapy in rehabilitation: a systematic review of the literature. Int.
J. Rehabil. Res. 34, 113.
Roux, F.E., Ibarrola, D., Lazorthes, Y., Berry, I., 2001. Virtual movements activate primary
sensorimotor areas in amputees: report of three cases. Neurosurgery 49, 736741
discussion 741-732.
Roux, F.E., Lotterie, J.A., Cassol, E., Lazorthes, Y., Sol, J.C., Berry, I., 2003. Cortical areas in-
volved in virtual movement of phantom limbs: comparison with normal subjects.
Neurosurgery 53, 13421352 discussion 1352-1343.
Roux, F.E., Ibarrola, D., Lazorthes, Y., Berry, I., 2008. Chronic motor cortex stimulation for
phantom limb pain: a functional magnetic resonance imaging study: technical case
report. Neurosurgery 62, 978985.
Saitoh, Y., Shibata, M., Sanada, Y., Mashimo, T., 1999. Motor cortex stimulation for phan-
tom limb pain. Lancet 353, 212.
Samii, A., Wassermann, E.M., Ikoma, K., Mercuri, B., Hallett, M., 1996. Characterization of
postexercise facilitation and depression of motor evoked potentials to transcranial
magnetic stimulation. Neurology 46, 13761382.
Samulack, D.D., Waters, R.S., Dykes, R.W., McKinley, P.A., 1990. Absence of responses to
microstimulation at the hand-face border in baboon primary motor cortex. Can.
J. Neurol. Sci. 17, 2429.
Schieber, M.H., Deuel, R.K., 1997. Primary motor cortex reorganization in a long-term
monkey amputee. Somatosens. Mot. Res. 14, 157167.
Schwenkreis, P., Witscher, K., Janssen, F., Pleger, B., Dertwinkel, R., Zenz, M., et al., 2001.
Assessment of reorganization in the sensorimotor cortex after upper limb amputa-
tion. Clin. Neurophysiol. 112, 627635.
Siegel, J.S., Power, J.D., Dubis, J.W., Vogel, A.C., Church, J.A., Schlaggar, B.L., Petersen, S.E.,
2014. Statistical improvements in functional magnetic resonance imaging analyses
produced by censoring high-motion data points. Hum. Brain Mapp. 35, 19811996.
Simoes, E.L., Bramati, I., Rodrigues, E., Franzoi, A., Moll, J., Lent, R., Tovar-Moll, F., 2012.
Functional expansion of sensorimotor representation and structural reorganization
of callosal connections in lower limb amputees. J. Neurosci. 32, 32113220.
Sun, Z.Y., Kloppel, S., Riviere, D., Perrot, M., Frackowiak, R., Siebner, H., Mangin, J.F., 2011.
The effect of handedness on the shape of the central sulcus. NeuroImage 60, 332339.
Topper, R., Foltys, H., Meister, I.G., Sparing, R., Boroojerdi, B., 2003. Repetitive transcranial
magnetic stimulation of the parietal cortex transiently ameliorates phantom limb
pain-like syndrome. Clin. Neurophysiol. 114, 15211530.
Trojan, J., Diers, M., Fuchs, X., Bach, F., Bekrater-Bodmann, R., Foell, J., Kamping, S., Rance,
M., Maass, H., Flor, H., 2014. An augmented reality home-training system based on
the mirror training and imagery approach. Behav. Res. Methods 46, 634640.
Tyc, F., Boyadjian, A., Allam, N., Brasil-Neto, J.P., 2012. Abnormal acute changes in upper
limb muscle cortical representation areas in the patients with writer's cramp during
co-activation of distal and proximal muscles. Acta Physiol (Oxf.) 206, 195207.
Tzourio-Mazoyer, N., et al., 2002. Automated anatomical labeling of activations in SPM
using a macroscopic anatomical parcellation of the MNI MRI single-subject brain.
NeuroImage 15, 273289.
Van Essen, D.C., Drury, H.A., Dickson, J., Harwell, J., Hanlon, D., Anderson, C.H., 2001. An
integrated software suite for surface-based analyses of cerebral cortex. J. Am. Med. In-
form. Assoc. 8, 443459.
Wu, C.W., Kaas, J.H., 1999. Reorganization in primary motor cortex of primates with long-
standing therapeutic amputations. J. Neurosci. 19, 76797697.
Xie, H., Kane, J.T., Dennis, M.J., Mooney, R.D., Bauer, W.R., Wang, X., Wall, J.T., 2013. Case
series evidence for changed interhemispheric relationships in cortical structure in
some amputees. J. Clin. Neurosci. 20, 523526.
Yang, T.T., Gallen, C.C., Ramachandran, V.S., Cobb, S., Schwartz, B.J., Bloom, F.E., 1994. Non-
invasive detection of cerebral plasticity in adult human somatosensory cortex.
Neuroreport 5, 701704.
Yao, J., Chen, A., Carmona, C., Dewald, J.P., 2009. Cortical overlap of joint representations
contributes to the loss of independent joint control following stroke. NeuroImage
45, 490499.