Seeing the phantom: a functional magnetic resonance imaging study of a supernumerary phantom limb.
ABSTRACT Supernumerary phantom limb (SPL) is a rare neurological manifestation where patients with a severe stroke-induced sensorimotor deficit experience the illusory presence of an extra limb that duplicates a real one. The illusion is most often experienced as a somesthetic phantom, but rarer SPLs may be intentionally triggered or seen. Here, we report the case of a left visual, tactile, and intentional SPL caused by right subcortical damage in a nondeluded woman.
Using functional magnetic resonance imaging, we investigated the multimodal nature of this phantom, which the patient claimed to be able see, use, and move intentionally. The patient participated in a series of sensorimotor and motor imagery tasks involving the right, the left plegic, and the SPL's hand.
Right premotor and motor regions were engaged when she imagined that she was scratching her left cheek with her left plegic hand, whereas when she performed the same task with the SPL, additional left middle occipital areas were recruited. Moreover, comparison of responses induced by left cheek (subjectively feasible) versus right cheek scratching (reportedly unfeasible movement) with the SPL demonstrated significant activation in right somesthetic areas.
These findings demonstrate that intentional movements of a seen and felt SPL activate premotor and motor areas together with visual and sensory cortex, confirming its multimodal dimension and the reliability of the patient's verbal reports. This observation, interpreted for cortical deafferentation/disconnection caused by subcortical brain damage, constitutes a new but theoretically predictable entity among disorders of bodily awareness.
- SourceAvailable from: Trevor J Dodds[Show abstract] [Hide abstract]
ABSTRACT: In neurology and psychiatry the detailed study of illusory own body perceptions has suggested close links between bodily processing and self-consciousness. One such illusory own body perception is heautoscopy where patients have the sensation of being reduplicated and to exist at two or even more locations. In previous experiments, using a video head-mounted display, self-location and self-identification were manipulated by applying conflicting visuo-tactile information. Yet the experienced singularity of the self was not affected, i.e., participants did not experience having multiple bodies or selves. In two experiments presented in this paper, we investigated self-location and self-identification while participants saw two virtual bodies (video-generated in study 1 and 3D computer generated in study 2) that were stroked either synchronously or asynchronously with their own body. In both experiments, we report that self-identification with two virtual bodies was stronger during synchronous stroking. Furthermore, in the video generated setup with synchronous stroking participants reported a greater feeling of having multiple bodies than in the control conditions. In study 1, but not in study 2, we report that self-location - measured by anterior posterior drift - was significantly shifted towards the two bodies in the synchronous condition only. Self-identification with two bodies, the sensation of having multiple bodies, and the changes in self-location show that the experienced singularity of the self can be studied experimentally. We discuss our data with respect to ownership for supernumerary hands and heautoscopy. We finally compare the effects of the video and 3D computer generated head-mounted display technology and discuss the possible benefits of using either technology to induce changes in illusory self-identification with a virtual body.Frontiers in Psychology 01/2013; 4:946. · 2.80 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Current neuropsychological evidence demonstrates that damage to sensory-specific and heteromodal areas of the brain not only disrupt the ability of combining sensory information from multiple sources, but can also cause altered multisensory experiences. On the other hand, there is also evidence of behavioural benefits induced by spared multisensory mechanisms. Thus, crossmodal plasticity can be viewed in both an adaptive and maladaptive context. The emerging view is that different crossmodal plastic changes can result following damage to sensory-specific and heteromodal areas, with post-injury crossmodal plasticity representing an attempt of a multisensory system to reconnect the various senses and by-pass injured areas. Changes can be considered adaptive when there is compensation for the lesion-induced sensory impairment. Conversely, it may prove maladaptive when atypical or even illusory multisensory experiences are generated as a result of re-arranged multisensory networks. This theoretical framework posits new intriguing questions for neuropsychological research and places greater emphasis on the study of multisensory phenomena within the context of damage to large-scale brain networks, rather than just focal damage alone.Neuroscience & Biobehavioral Reviews 12/2012; · 9.44 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Supernumerary phantom limb (SPL) resulting from spinal cord lesions are very rare, with only sporadic and brief descriptions in the literature. Furthermore, the reported cases of SPL typically occurred in neurologically incomplete spinal cord patients. Here, we report a rare case of SPL with phantom limb pain that occurred after traumatic spinal cord injury in a neurologically complete patient. After a traffic accident, a 43-year-old man suffered a complete spinal cord injury with a C6 neurologic level of injury. SPL and associated phantom limb pain occurred 6 days after trauma onset. The patient felt the presence of an additional pair of legs that originated at the hip joints and extended medially, at equal lengths to the paralyzed legs. The intensity of SPL and associated phantom limb pain subsequently decreased after visual-tactile stimulation treatment, in which the patient visually identified the paralyzed limbs and then gently tapped them with a wooden stick. This improvement continued over the 2 months of inpatient treatment at our hospital and the presence of the SPLs was reduced to 20% of the real paralyzed legs. This is the first comprehensive report on SPLs of the lower extremities after neurologically complete spinal cord injury.Annals of rehabilitation medicine. 12/2013; 37(6):901-6.
Seeing the Phantom: A Functional Magnetic
Resonance Imaging Study of a
Supernumerary Phantom Limb
Asaid Khateb, PhD,1–3Ste ´phane R. Simon, PhD,4Sebastian Dieguez, MSc,5Franc ¸ois Lazeyras, PhD,4
Isabelle Momjian-Mayor, MD,2Olaf Blanke, Pr MD,2,5Theodor Landis, Pr MD,2,3Alan J. Pegna, PhD,1–3and
Jean-Marie Annoni, Pr MD2,3
Objective: Supernumerary phantom limb (SPL) is a rare neurological manifestation where patients with a severe stroke-induced
sensorimotor deficit experience the illusory presence of an extra limb that duplicates a real one. The illusion is most often
experienced as a somesthetic phantom, but rarer SPLs may be intentionally triggered or seen. Here, we report the case of a left
visual, tactile, and intentional SPL caused by right subcortical damage in a nondeluded woman.
Methods: Using functional magnetic resonance imaging, we investigated the multimodal nature of this phantom, which the
patient claimed to be able see, use, and move intentionally. The patient participated in a series of sensorimotor and motor
imagery tasks involving the right, the left plegic, and the SPL’s hand.
Results: Right premotor and motor regions were engaged when she imagined that she was scratching her left cheek with her
left plegic hand, whereas when she performed the same task with the SPL, additional left middle occipital areas were recruited.
Moreover, comparison of responses induced by left cheek (subjectively feasible) versus right cheek scratching (reportedly unfea-
sible movement) with the SPL demonstrated significant activation in right somesthetic areas.
Interpretation: These findings demonstrate that intentional movements of a seen and felt SPL activate premotor and motor
areas together with visual and sensory cortex, confirming its multimodal dimension and the reliability of the patient’s verbal
reports. This observation, interpreted for cortical deafferentation/disconnection caused by subcortical brain damage, constitutes
a new but theoretically predictable entity among disorders of bodily awareness.
Ann Neurol 2009;65:698–705
Phantom limbs are usually described after amputation
of extremities and can be defined as “the persistent ex-
perience of the postural and motor aspects of a limb
after its physical loss.”1In hemiplegic patients after
stroke, a similar phenomenon can occasionally occur,
consisting in the experience of an additional limb.2–5
Much less frequent than amputation phantoms, this
phenomenon, generally called supernumerary phantom
limb (SPL), refers to “the awareness of having an extra
limb in addition to the regular set of two arms and two
Compared with postamputation phantoms,
which are overwhelmingly criticized by the patient as
an illusory feeling, SPLs are sometimes part of delu-
sional syndromes involving reduplications of body
parts or “pseudopolymyelia.”6On other instances, the
phenomenon can be more selective and the patient
may retain a critical attitude toward his experience of
an “extra limb.” These generally involve focal subcor-
tical lesions, and must be differentiated from patients
affected by multiple delusional reduplications and suf-
fering from larger cortical predominantly right lesion.7
Phenomenologically, the literature suggests a large
spectrum of features. In most cases, the SPL has only
somesthetic characteristics and cannot be seen or pur-
SPLs with motor components
mostly involve anarchic, uncontrollable movements,9
very restricted ones,10or mirror automatically the
healthy limb’s movements.11,12To our knowledge,
only one case with a highly controllable and purposeful
SPL has been reported and described as an “intentional
motor phantom limb.”13There are scattered reports of
involvement of other modalities concomitantly to a
From the1Laboratory of Experimental Neuropsychology and2De-
partment of Neurology, Geneva University Hospitals and Faculty of
4Center for Biomedical Imaging and Department of Radiology, Ge-
neva University Hospitals and Faculty of Medicine, Geneva; and
5Laboratory of Cognitive Neuroscience, Brain Mind Institute, Ecole
Polytechnique Fe ´de ´rale de Lausanne, Lausanne, Switzerland.
3Faculty of Medicine, Geneva Neuroscience Centre;
Address correspondence to Dr Khateb, Laboratory of Experimental
Neuropsychology, Geneva University Hospitals, 24, rue Micheli-du-
Crest, 1211 Geneva 14, Switzerland. E-mail: email@example.com
Potential conflict of interest: Nothing to report.
Additional Supporting Information may be found in the online ver-
sion of this article.
Received Sep 2, 2008, and in revised form Dec 19. Accepted for
publication Jan 16, 2009.
Published in Wiley InterScience (www.interscience.wiley.com).
© 2009 American Neurological Association
somesthetic SPL in the literature. A few patients re-
ported a visual component to their SPL, saying they
could see it.2,4,14–19Still other cases were described
with tactile components, with patients reporting that
they can actually feel objects or body parts with their
phantom.12,14,20–23These intentional and visual as-
pects, although rare and poorly detailed, indicate that
multimodal mechanisms might be involved in SPLs.
Given the abundance of components that might
constitute the phenomenology of SPLs, different neural
mechanisms have been proposed. Classic principles em-
phasize a dissociation between the previously estab-
lished sensorimotor representations and the lesion-
induced change in communication between the brain
and the paralyzed limb (see Brugger’s article5). A defi-
cit in the representation of space,6disordered reality-
testing abilities,2,24or failed updating of motor com-
mands25have also been proposed for SPLs.
Of special interest for the comprehension of the
brain mechanisms of SPL are case studies using func-
tional magnetic resonance imaging (fMRI). A recent
study assessing the neural correlates of intentional
movements of an SPL in a patient with subcortical
stroke suggested an abnormal reorganization within the
motor system. This hypothesis was based on the obser-
vation that SPL’s movements were associated with in-
creased fMRI responses in motor areas and in the bi-
lateral basal ganglia–thalamus.13The authors propose
that cortical pathways during such movements “may
reflect an abnormal closed-loop functioning of the
thalamocortical system” caused by the interruption of
thalamic sensory afferences.13In another patient expe-
riencing an SPL after a frontomesial stroke, the motor
system was implicated in the illusion but at a more
integrative level.26Functional activations were reported
in the right internal supplementary motor area during
the presence of the SPL. The authors thus proposed
that an uncontrolled reactivation of motor areas during
self-triggered voluntary action might influence the per-
ceptual signals from the arm and cause the SPL. Taken
together, such views favor the hypothesis of an inter-
action between motor and somesthetic processes in
If a polymodal perceptive interaction has to be re-
tained, it is difficult to ignore a possible role of the
visual system in the elicitation of this perceptive mis-
match. In certain patients, vision can cancel phantom
limbs or their associated features such as pain.27Ac-
cordingly, one can reasonably assume that an aberrant
visual coactivation may influence motor or polymodal
areas, and thus play a role in the appearance of an SPL.
Here, we report a clinical and fMRI investigation of a
patient who suffered from a right subcortical stroke
and experienced a SPL that she could feel, see, and
intentionally move. Our aim was to assess the neural
basis of this multimodal phenomenon using motor and
sensorimotor tasks involving the SPL. According to the
patient’s reports, we asked her to scratch her right or
left cheek with the SPL during functional imaging ses-
sions. We hypothesized that scratching would require
that she bring her SPL close to her eyes and, therefore,
produce task-related activations in the visual and sen-
sory cortices, reflecting the reportedly vivid phenome-
nal experience of her SPL.
Subjects and Methods
The patient, a 64-year-old librarian with a medical history of
treated hypertension and occasional depression, was admitted
to the emergency department in March 2007 after a sudden
onset of left-sided weakness and dysarthria. She had appar-
ently driven a long distance in a hemiplegic state, and it was
only on arrival at her destination, as she tried to walk, that
her hemiplegia was noticed by a passerby, indicating the
presence of anosognosia at least initially. At admission, her
neurological evaluation showed an National Institutes of
Health Stroke Scale score of 16. She was well oriented in
time and place, cooperative, presented mild spatial neglect,
initial left hemianopia, left facial and leg paresis, left arm
plegia, hypoesthesia to pain, and severely decreased positional
sense. The computed tomographic scan and MRI (see Re-
sults) confirmed a subcortical capsulolenticular hemorrhage.
Behavioral changes included misoplegia, increased impulsiv-
ity, and emotional swing, as well as urinary incontinence
during at least the first month of her hospitalization. Formal
neuropsychological examination at 1 week (see Supplemen-
tary Table S1) excluded diffuse cognitive alteration and con-
firmed the presence of a left spatial neglect without signs of
personal neglect. At this point, anosognosia had entirely re-
The intriguing symptom was the subjective appearance,
from the 4th day on, of an SPL, starting from the elbow of
the left plegic arm, which the patient claimed she could
move, see, and even use to touch parts of her body (Fig 1;
see supplementary materials for more details). Her SPL was
not experienced permanently, but only when she decided to
“trigger” it intentionally, thus meeting the rare defining fea-
ture of the “intentional motor phantom limb syndrome.”13
The SPL was visually perceived whenever it was “triggered.”
She described it as “pale,” “milk-white,” and “transparent.”
Tactile sensations in the SPL happened when she clenched
her hand (she could then feel her phantom palm with her
phantom fingers) and when she “touched” certain parts of
her body (in which case, the sensation was felt both in the
phantom and the touched body part). She could touch parts
of her head, as well as her right shoulder. She claimed to be
able to use the SPL to scratch an itch on her head (with an
actual sense of relief). Moreover, she reported that the phan-
tom could not penetrate solid obstacles (see supplementary
materials for more details).
This phenomenon persisted during the following 4
months spent in the neurorehabilitation clinic and decreased
progressively over the next few months, as she regained pro-
prioceptive feelings and motor abilities in the plegic limb. At
discharge, spatial neglect had completely receded, hemi-
Khateb et al: A Phantom Limb Studied by fMRI
anopia turned into a quadrantopia, and hemiplegia partially
recovered, particularly for the lower left limb.
Functional Magnetic Resonance Imaging Investigation
At 3 weeks after stroke, this study was designed to investigate
the polymodal nature of the SPL. Our aim was to compare
fMRI responses induced by real movements (RM) of the
healthy right hand, imaginary movements (IM) of the right
and left hands, and movements of the SPL, thus using the
patient as her own control across conditions. Based on the
patient’s claim that she could see, move, and use her SPL, we
predicted that SPL’s movements might elicit activations in
motor, somatosensory, and visual areas (see the paradigm de-
tailed in the supplementary materials).
The formal neuropsychological assessment conducted 1
week after stroke showed mainly cognitive difficulties
related to spatial neglect (see Supplementary Table S1).
With particular relevance to our study, when ques-
tioned before the fMRI session, the patient still expe-
rienced her SPL, with no changes relative to previous
clinical observations. After each MRI scan, the patient
was systematically questioned about the feasibility of
the task she had just performed (whether she saw and
felt the SPL across conditions). Globally, she declared
that she had performed the different tasks correctly.
Thus, she could perform RM with the right hand and
IM with both. When asked to perform movements
with the SPL, she explained that, although she was able
to scratch her left cheek correctly, she experienced
some difficulties in scratching the right one because of
the fact that the narrow space left by the scanner an-
tenna “prevented” her SPL from reaching the right
cheek, thus confirming her previous report that she was
unable to move her phantom through obstacles.
SIMPLE MOTOR TASKS.
The fMRI results of simple motor tasks conducted
with the healthy right and the paretic hand are sum-
marized in Figures 2B to D. The RMs of the right
hand were, as expected, associated with dominant acti-
vation of the left sensorimotor (M1, S1) and premotor
cortex (ventrolateral, dorsolateral, as well as medial pre-
motor areas), left temporooccipital junction, left supe-
rior parietal lobule, and possibly a small contralateral
hand representational cortex (see Fig 2B). Similar but
less extensive activations were induced by the IMs of
the same hand (see Fig 2C). More interestingly, the
IMs of the paralyzed left hand activated the right mo-
tor and the bilateral dorsal premotor cortex, the sup-
plementary motor area, and the left motor cortex (see
Fig 2D and details in Supplementary Table S2).
Figure 2E illustrates the activation induced by the
scratching of the right cheek by the SPL. Although the
patient stated that she was unable to perform these
movements correctly because of the scanner antenna,
the activation involved the left and right hand/arm ar-
eas (M1), the premotor cortex, but also the left occip-
ital cortex (BA 19/18, see details in Supplementary Ta-
ble S2), compatible with the patient’s experienced
sensation of “moving and seeing” her SPL. Scratching
the left cheek with the SPL (see Fig 2F), a task that she
said was performed correctly, activated almost exactly
the same premotor regions, the left occipital cortex,
but here more extensively the right hand/arm M1 (as
well as other areas detailed in the Table), again con-
cordant with her report that she “moved and saw” her
SPL. In contrast, scratching this same left cheek with
the paretic left hand (see Fig 2G) activated the right
motor and bilateral premotor areas (see the Table), but
not the middle occipital areas.
Fig 1. Schematic representation of the phantom limb as
drawn by the patient. Note the left-right reversal of the body
and the supernumerary phantom limb (SPL) emerging at the
level of the elbow of the paralyzed arm, indicated here by the
French word inutile (for “useless”; see additional notes in sup-
Annals of Neurology Vol 65 No 6 June 2009
Fig 2. (A) Magnetic resonance imaging axial, sagittal, and coronal images showing the localization of the subcortical lesion. (B–D)
Activation maps obtained respectively during real right hand, imaginary right, and left plegic hand movements. (E, F) Activation
maps for the supernumerary third hand movements scratching the right (E) and the left cheek (F). (G) Imaginary left hand move-
ments scratching the left cheek. (H) Comparison of left cheek scratching by the third phantom hand versus the left paralyzed hand
(contrast illustrated in E minus G). (I) Comparison of the left cheek scratching versus the right cheek scratching with the third
hand. (J) Parameter estimates of activation for two regions of interest over the different conditions illustrated in (B–G). (J, left)
Parameter estimates for the middle occipital gyrus found mainly in (E) and (F) in relation to the phantom movements. (J, right)
Parameter estimates for the somesthetic S1 area demonstrated by the contrast in (I) (contrast F minus E). [Color figure can be
viewed in the online issue, which is available at www.interscience.wiley.com.]
Khateb et al: A Phantom Limb Studied by fMRI
Table. Activated Regions with Their Talairach x, y, z Coordinates, Maximum Value of the t-Statistics of the
Reported Cluster, the Cluster Sizes (K), and the Corresponding Brodmann Areas in the Sensorimotor Tasks
Anatomic Location x, y, zTmax
SPL Scratching Left Cheeka
Right M1 (hand/arm areas)
Left M1 (hand area)
Left dorsolateral premotor cortex
Right ventrolateral premotor cortex
Left lingual gyrus
Left middle occipital gyrus
Left superior temporal gyrus
Left temporoparietooccipital junction
Left inferior frontal gyrus
Left Paretic Hand Scratching Left Cheekb
Right M1 (hand/arm areas)
Right M1 (arm area)
Right dorsolateral premotor cortex
Left dorsolateral premotor cortex
Left temporoparietooccipital junction
Right superior parietal lobule
Left Cheek Scratching: SPL versus Paretic Left Handc
Left lingual gyrus
Left middle occipital gyrus
Left middle frontal gyrus
Left inferior frontal gyrus
SPL: Scratching Left versus Right Cheekd
Right somesthetic cortex (face area S1)
31, ?22, 58
23, ?35, 62
?34, ?29, 49
?11, ?15, 64
4, ?14, 61
?45, ?10, 45
?47, 1, 26
55, 0, 12
?14, ?57, ?2
?24, ?96, 13
?50, ?49, 23
?41, ?76, 5
?47, 19, 9
31, ?23, 58
18, ?35, 65
37, ?1, 48
?35, ?9, 42
?43, ?70, 8
31, ?51, 55
?13, ?54, ?1
?5, ?77, 14
?14, ?88, 14
?40, 46, 9
?47, 23, 9
47, ?27, 44
43, ?34, 28
aActivated brain regions induced by the “scratching” of the left cheek with the supernumerary phantom limb (SPL).bActivated regions
induced by “scratching” the left cheek with the plegic left hand.cSpecific activation demonstrated by the comparison of left cheek
scratching movements with the SPL versus the left plegic hand (ie, contrast the first section of the table minus the second section).
dActivation of somesthetic areas demonstrated by the comparison of the left cheek scratching (said to be feasible by the patient) versus
the right cheek (said to be unfeasible because of the magnetic resonance imaging scanner antenna preventing her to reach her cheek)
with the SPL. Note that the first and second sections of the table refer to activation obtained from direct contrast of activation versus
resting condition, whereas the third and fourth sections of the table refer to activation obtained from interaction analysis between two
different contrasts (see details in supplementary materials). BA ? Brodmann area; SMA ? supplementary motor area.
Annals of NeurologyVol 65No 6June 2009
To further substantiate this later observation and as-
sess the involvement of sensory cortex during scratch-
ing tasks, we conducted interaction analyses between
the contrasts. The specificity of the occipital involve-
ment during SPL movements was assessed by the com-
parison of activation induced by the scratching of the
left cheek with the SPL versus the plegic hand (see Fig
2H). This showed mainly significant voxels in the left
middle occipital (BA 18/19; see the Table) and less ex-
tensively in prefrontal areas. Figure 2J (left), which il-
lustrates the parameter estimates of activation for the
middle occipital area, confirms that activity in this re-
gion was specifically related to SPL movements. Fi-
nally, the interaction between the scratching of the left
cheek (declared as possible) versus the right cheek (im-
possible) by the SPL showed significant activation only
in the right somesthetic S1 face area (see Fig 2I; see the
Table). The parameter estimates displayed in Figure 2J
(right) confirm that area S1 displayed the highest acti-
vation during left cheek scratching (see Fig 2J, lane F).
This measurable sensory response suggests that the pa-
tient, indeed, experienced the sensation of left cheek
scratching. However, although the “phantom touch”
was reported as a “double touch” (ie, sensations both
on SPL and the touched body part), the responses to
such double-touch feelings are indistinguishable here,
presumably because of the proximity of the hand and
face receptive fields.
This article describes the clinical features and neural
correlates of a multimodal poststroke SPL characterized
by the presence of motor, tactile, and visual compo-
nents. The evidence presented here demonstrates a di-
rect link between the sensorimotor characteristics of
the illusion and modality-specific activations in motor,
visual, and somesthetic areas. Although previous re-
ports have described various clinical features of SPLs,
few studies investigated the neural correlates of this
phenomenon. In this respect, the case presented here
is, to our knowledge, unique for its multimodal char-
acteristics. Furthermore, this functional investigation
appears to be the first to assess the neural basis of a
multimodal stroke-induced SPL.
The visual characteristics of SPLs have not been dis-
cussed much in the stroke literature because the major-
ity of such patients do not report seeing the phantom
(nine visible SPL cases reported; see earlier). Even in
the only other report of intentional motor SPL, the
patient could move but not see the limb.13Some SPLs
are reported as not seen and not influenced, either by
visual feedback of the actual limbs or direct touch of
the actual (and reduplicated) limb.6,13In the few cases
of nonintentional SPLs, visibility was reported, al-
though less precisely documented than here, and
mostly in a delusional frame of mind. Inversely, in
some instances, the SPL could even be cancelled by
vision and tactile stimulation of the real limb.28The
presence of multimodal components in some cases sug-
gests that these illusory phenomena might be better ex-
plained by a failure to correctly integrate multimodal
and motor information to generate a normal experi-
ence of self-location29,30than by sensorimotor reorga-
nization as in amputees,31although these two ap-
proaches are certainly not exclusive. The literature
indeed suggests the existence of various types of SPLs
differing according to the phenomenological richness
(unimodal, bimodal, multimodal), degree of motor
movements, and so forth), sensitivity to sensory feed-
back (visual, tactile, proprioceptive), and patients’
mental framework (confused, delusional, indifferent,
lucid). In this respect, it appears unreasonable to expect
a single explanation for such a clinically complex and
Although rare, previous studies have assessed the
neural correlates of the presence, or of the intentional
movements, of stroke-induced SPL,13,26,28but this in-
vestigation is the first to demonstrate the multimodal
aspects of this phenomenon. Particularly, visual activa-
tion accompanying SPL movements has never been ob-
served either in amputation32or in poststroke phan-
toms.13,26Also, the activation of somesthetic areas fits
with the patient’s reported conscious experience of feel-
ing and seeing her SPL, thus reinforcing our opinion
that she was not simply confabulating.
The activation of the left extrastriate areas involving
the occipito-temporo-parietal junction appears explain-
able. This area has often been involved in the alteration
of body schema perception.33–36Its electrical stimula-
tion was also found to induce illusory perception (pro-
prioceptive and visual) of body part movements.34
Functional studies have also showed that the “extrastri-
ate body area” responds to the perception of other peo-
ple’s body parts and to goal-directed movements (or
imagined movements) of own body parts.37,38Here,
the activation coordinates correspond to those previ-
ously reported for the extrastriate body area (x, y, z ?
?45, ?65, 2 in Peelen and Downing’s article39; x, y,
z ? ?48; ?69;6 in Astafiev and colleagues’ article37).
In our study, although activation in this area was sig-
nificant in some contrasts only (including RM and IM
of the right hand; see Figs 2B, C), the analysis of the
parameter estimates (not illustrated) showed its recruit-
ment in all conditions and not only in relation to the
SPL. Additional analysis in two control subjects
showed the bilateral involvement of this area during
RM of the right and left hand (unpublished data). Ac-
cordingly, we hypothesize that the perception of the
flickering hand during the activation condition (vs the
fixation during the resting condition) is the most plau-
Khateb et al: A Phantom Limb Studied by fMRI
sible explanation for the involvement of this left tem-
However, the activation found in the left occipital
lobe was specifically related to the SPL movements,
and involved several medial-posterior foci extending
beyond the primary visual area V1, mostly area V2.40
Activity in such low-level visual areas (including the
lingual gyrus and the cuneus) has been observed during
visually guided eye and limb movements,37but also in
various other contexts. Thus, these areas were activated
by motor imagery tasks,41mental generation of words
or objects images,42–44and illusory contours.45In light
of these observations supporting the participation of
low-level visual areas in contexts implying goal-
oriented movements, motor and mental imagery, the
activation discussed here during “visualization” of the
SPL’s intentionally triggered and goal-directed move-
ments appears to make sense. Finally, though the acti-
vation of these areas involved the bilateral visual cortex
in control subjects,37,39the lack of right activation in
our patient might reasonably be explained by the initial
hemianopia, possibly because of a hemorrhage-induced
disconnection of these areas.
The anatomical disruption and the ensuing somes-
thetic deafferentation in our and other patients consti-
tutes a necessary but not a sufficient condition for the
appearance of an SPL. Existing evidence from stroke-
elicited SPLs convincingly implicates the mismatch be-
tween the subject’s well-established sensorimotor repre-
communication between the brain and the paralyzed
limb. In this respect, understanding of the phenome-
nological complexity of SPLs and the puzzles they pose
could benefit from the examination of the general
framework of the body schema alterations (eg, somato-
paraphrenia, hemiasomatognosia, out-of-body experi-
ence, feeling of a presence, autoscopy, and heutoscopy;
see Arzy and coauthors’33and Blanke and researchers’
articles34). This view is concordant with Todd and
Dewhurst’s19early statement that “the supernumerary
phantom arm is in the nature of an autoscopic double of
the real arm.” However, given the unique characteris-
tics of our patient’s SPL, we are tempted to interpret it
as a new neurological syndrome, filling the gap be-
tween unseen duplicated body parts in the peripersonal
space and visible doubles of one’s entire body in the
extrapersonal space. In other words, and in line with
other authors’ views arguing for the continuity of dis-
turbances from body parts to whole-body illusions,46,47
this case might constitute, at least conceptually and
phenomenologically, a missing link between classic
phantom limbs and autoscopic phenomena. Yet, the
question remains why these two well-known phenom-
ena are relatively frequent in nondelusional patients,
whereas our patient’s visible SPL is so unusual. More-
over, the questions of how disorders of the body
aberrant pattern of
schema, extending from isolated body parts to the en-
tire body, are linked at the neurofunctional level and
whether such disorders can ultimately fit a single ex-
planatory model also remain moot points. Ultimately,
however, these conditions might offer a unique way to
understand how the brain constructs a normal experi-
ence of bodily awareness and the self.
This research was supported by the Swiss National Science Founda-
tion (325100-118362) and the Center for Biomedical Imaging of
Geneva and Lausanne (S.S.).
This article is dedicated to the memory of our friend
and colleague neuropsychologist Euge `ne Mayer (James)
who tragically passed away on December 14, 2008. We
thank F. Henry for technical assistance.
1. Brugger P. From phantom limb to phantom body. In:
Knoblich G, Thornton IM, Grosjean M, Shiffrar M, eds. Hu-
man body perception from the inside out. Oxford: Oxford Uni-
versity Press, 2006:171–209.
2. Halligan PW, Marshall JC, Wade DT. Three arms: a case study
of supernumerary phantom limb after right hemisphere stroke.
J Neurol Neurosurg Psychiatry 1993;56:159–166.
3. Weinstein EA, Kahn RL, Malitz S, Rozanski J. Delusional re-
duplication of parts of the body. Brain 1954;77:45–60.
4. Halligan PW, Marshall JC. Supernumerary phantom limb after
right hemispheric stroke. J Neurol Neurosurg Psychiatry 1995;
5. Brugger P. Supernumerary phantoms: a comment on Grossi, et
al.’s (2002) spare thoughts on spare limbs. Percept Mot Skills
6. Vuilleumier P, Reverdin A, Landis T. Four legs. Illusory redu-
plication of the lower limbs after bilateral parietal lobe damage.
Arch Neurol 1997;54:1543–1547.
7. Frederiks JA. Occurrence and nature of phantom limb phe-
nomena following amputation of body parts and following le-
sions of the central and peripheral nervous system. Psychiatr
Neurol Neurochir 1963;66:73–97.
8. Grossi D, Di Cesare G, Tamburro RP. On the syndrome of the
“spare limb”: one case. Percept Mot Skills 2002;94:476–478.
9. van Bogaert L. Sur la pathologie de l’image de soi. Ann Med
Psychol 1934;92:519–555; 746–759.
10. Pine ´as H. Ein Fall von phantoma ¨hnlichen Erscheinungen
(“Phantomarm”) bei hemiplegischer La ¨hmung. Nervenarzt
11. Bakheit AM. Phantom limb sensations after complete thoracic
transverse myelitis. J Neurol Neurosurg Psychiatry 2000;69:
12. Lhermitte F, Cambier J. Les perturbations somatognosiques en
pathologie nerveuse. Paris: Masson, 1960:465–554.
13. Staub F, Bogousslavsky J, Maeder P, et al. Intentional motor
phantom limb syndrome. Neurology 2006;67:2140–2146.
14. Bakheit AM, Roundhill S. Supernumerary phantom limb after
stroke. Postgrad Med J 2005;81:e2.
15. Critchley M. The parietal lobes. London: Edward Arnold,
16. He ´caen H, de Ajuriaguerra J, Le Guillant L, Angelergues R.
De ´lire centre ´ sur un membre fanto ˆme chez un he ´miple ´gique
gauche par le ´sion vasculaire avec anosognosie. Evol Psychiatr
Annals of Neurology Vol 65No 6June 2009
17. Lenz H. Anfa ¨lle von Verdopplung der rechten Ko ¨rperha ¨lfte
sowie Verzerrung des Fu ¨hl- und Schaubildes rechts. Nervenarzt
18. Miyazawa N, Hayashi M, Komiya K, Akiyama I. Supernumer-
ary phantom limbs associated with left hemispheric stroke: case
report and review of the literature. Neurosurgery 2004;54:
19. Todd J, Dewhurst K. The double: its psychopathology and psy-
chophysiology. J Nerv Ment Dis 1955;122:47–55.
20. Bakheit AM. Illusory movements of the paralysed upper limb
in stroke. J Neurol Neurosurg Psychiatry 2000;69:837.
21. He ´caen H, de Ajuriaguerra J. Me ´connaissances et hallucinations
corporelles. Paris: Masson, 1952.
22. Lohmann AM. Ein Fall von Phantomerscheinungen bei einer
linksseitigen Hemiplegie. Dtsch. Z. Nervenheilk, 1935:136.
23. Mazzoni M, Lucchini C, Meucci G, et al. Supernumerary pi-
anto limb after ischaemic stroke. Neurocase 1997;3:223–230.
24. Sellal F, Renaseau-leclerc C, Labrecque R. [The man with 6
arms. An analysis of supernumerary phantom limbs after right
hemisphere stroke]. Rev Neurol (Paris) 1996;152:190–195.
25. Frith CD, Blakemore SJ, Wolpert DM. Abnormalities in the
awareness and control of action. Philos Trans R Soc Lond B
Biol Sci 2000;355:1771–1788.
26. McGonigle DJ, Hanninen R, Salenius S, et al. Whose arm is it
anyway? An fMRI case study of supernumerary phantom limb.
27. Giraux P, Sirigu A. Illusory movements of the paralyzed limb
restore motor cortex activity. Neuroimage 2003;20(suppl 1):
28. Hari R, Hanninen R, Makinen T, et al. Three hands: fragmen-
tation of human bodily awareness. Neurosci Lett 1998;240:
29. Blanke O, Landis T, Spinelli L, Seeck M. Out-of-body experi-
ence and autoscopy of neurological origin. Brain 2004;127:
30. Brugger P. Reflective mirrors: perspective-taking in autoscopic
phenomena. Cogn Neuropsychiatry 2002;7:179–194.
31. Ramachandran VS, Rogers-Ramachandran D. Phantom limbs
and neural plasticity. Arch Neurol 2000;57:317–320.
32. Roux FE, Lotterie JA, Cassol E, et al. Cortical areas involved in
virtual movement of phantom limbs: comparison with normal
subjects. Neurosurgery 2003;53:1342–1353.
33. Arzy S, Seeck M, Ortigue S, et al. Induction of an illusory
shadow person. Nature 2006;443:287.
34. Blanke O, Ortigue S, Landis T, Seeck M. Stimulating illusory
own-body perceptions. Nature 2002;419:269–270.
35. Coslett HB. Evidence for a disturbance of the body schema in
neglect. Brain Cogn 1998;37:527–544.
36. Halligan PW, Marshall JC, Wade DT. Unilateral somatopara-
phrenia after right hemisphere stroke: a case description. Cortex
37. Astafiev SV, Stanley CM, Shulman GL, Corbetta M. Extrastri-
ate body area in human occipital cortex responds to the perfor-
mance of motor actions. Nat Neurosci 2004;7:542–548.
38. Downing PE, Jiang Y, Shuman M, Kanwisher N. A cortical
area selective for visual processing of the human body. Science
39. Peelen MV, Downing PE. Selectivity for the human body in
the fusiform gyrus. J Neurophysiol 2005;93:603–608.
40. Tootell RB, Hadjikhani NK, Vanduffel W, et al. Functional
analysis of primary visual cortex (V1) in humans. Proc Natl
Acad Sci U S A 1998;95:811–817.
41. Bakker M, De Lange FP, Helmich RC, et al. Cerebral corre-
lates of motor imagery of normal and precision gait. Neuroim-
42. Kosslyn SM, Thompson WL, Kim IJ, Alpert NM. Topograph-
ical representations of mental images in primary visual cortex.
43. D’Esposito M, Detre JA, Aguirre GK, et al. A functional MRI
study of mental image generation. Neuropsychologia 1997;35:
44. Slotnick SD, Thompson WL, Kosslyn SM. Visual mental im-
agery induces retinotopically organized activation of early visual
areas. Cereb Cortex 2005;15:1570–1583.
45. Seghier ML, Vuilleumier P. Functional neuroimaging findings
on the human perception of illusory contours. Neurosci Biobe-
hav Rev 2006;30:595–612.
46. Wade NJ. Beyond body experiences: phantom limbs, pain and
the locus of sensation. Cortex 2008;5:5.
47. Brugger P, Regard M, Landis T. Illusory reduplication of one’s
own body: phenomenology and classification of autoscopic phe-
nomena. Cogn Neuropsychiatry 1997;1:19–38.
Khateb et al: A Phantom Limb Studied by fMRI