Neuroscience Letters 240 (1998) 131–134

Three hands: fragmentation of human bodily awareness
Riitta Hari a , c ,*, Ritva Hänninen b, Timo Mäkinen b, Veikko Jousmäki a, Nina Forss a,
Mika Seppä a, Oili Salonen d
a

Brain Research Unit, Low Temperature Laboratory, Helsinki University of Technology, P.O. Box 2200, FIN-02015 HUT, Espoo, Finland
b
Department of Neurology, Central Hospital of Central Finland, Keskussairaalantie 19, FIN-40620 Jyväskylä, Finland
c
Department of Clinical Neurosciences, Helsinki University Central Hospital, FIN-00290 Helsinki, Finland
d
Department of Radiology, Helsinki University Central Hospital, FIN-00290 Helsinki, Finland
Received 5 September 1997; received in revised form 5 December 1997; accepted 5 December 1997

Abstract
We describe patient E.P. who occasionally perceives a ‘ghost’ hand which copies the previous positions of the left hand with a
0.5–1 min time lag, but follows the movement patterns of the right hand. The symptoms started after an operation of a ruptured
aneurysm, followed by an infarction of the right frontal lobe; E.P. also has a previously lesioned corpus callosum. Neuromagnetic
recordings revealed that activity of the left secondary somatosensory cortex was strongly suppressed during the ghost arm
percept, thereby providing an objective correlate for E.P.’s sensations. We conclude that simultaneous mental contents about
body scheme may be based on neural information extracted at considerably different times, resulting in fragmentation of bodily
awareness.  1998 Elsevier Science Ireland Ltd.

Keywords: Body image; Bodily awareness; Anterior supplementary motor area; Corpus callosum; Secondary somatosensory
cortex; Magnetoencephalography

A healthy human’s unitary bodily awareness relies on
a stable body scheme, which may be in part genetically
determined and which may change after brain lesion. We
report here on a patient whose bodily awareness became
fragmented after operation of a ruptured aneurysm of the
left pericallosal artery, resulting in a frontal lobe lesion.
E.P., a previously healthy 37-year-old right-handed
female suffered from subarachnoidal hemorrhage in January 1994. She had no psychiatric history and had worked
successfully before her illness in a paper product factory.
The ruptured aneurysm of the left pericallosal artery was
operated on but the recovery was slow, in part because of an
infarction in the right dorsomedial frontal lobe (Fig. 1). The
MRI also showed an apparently prenatal lesion of the body
of corpus callosum.
According to postoperative tests, E.P. had average intelligence and a very good memory. Her left hand was slightly
clumsy, easily started to follow the right hand, and was
* Corresponding author. Fax: +358 9 4512969;
e-mail: hari@neuro.hut.fi

disturbed during rhythmical movements. E.P. had slight
tactile anomia in the left hand and left-hand dominant constructional apraxia, but she was able to write and draw with
both hands. As a sign of motor perseveration she had difficulties stopping some rhythmical movements. When
attempting to greet with the right hand she sometimes raised
both hands. Position sense was normal. E.P.’s facial expression was often rigid when she was been spoken to or when
she did not talk herself. She was not able to change her
mind, plans, or order of actions quickly, and she often had
competing thoughts (e.g. ‘to go in or to continue walking
outside’) which paralyzed her.
After the operation E.P. had suffered from intermanual
conflict and alien hand syndrome, in agreement with her
partial callosal disconnection and the infarction of the
mesial frontal cortex [2,3]. For example, while E.P. was
reading a newspaper and turning the page with the right
hand, the left hand turned back the same page. While eating,
the left hand tried to put more food into the mouth than what
E.P. was able to swallow. During swimming, E.P. was
afraid that the left hand wanted to drown her.

0304-3940/98/$19.00  1998 Elsevier Science Ireland Ltd. All rights reserved
PII S0304- 3940(97) 00945- 2

132

R. Hari et al. / Neuroscience Letters 240 (1998) 131–134

Fig. 1. Top: surface reconstruction of E.P.’s brain based on magnetic
resonance images (Siemens 1.5-T Magnetomy) recorded 3 years
after the operation. For clarity, the infarcted right frontal region has
been removed by the computer. Bottom: the arrows point to lesions
of the right mesial frontal lobe, the anterior corpus callosum, and the
evidently perinatal lesion of the body of corpus callosum.

E.P. also developed other disturbances in her body image
[9]. Several times a day she sensed a third arm and, less
frequently, a third leg. Fig. 2 illustrates her own drawing of
such a situation (the left/right mistake in the drawing was
frequent for her). E.P. was fully conscious of the abnormality of her percepts and was able to observe analytically her
symptoms. The ghost limbs were always on the left side of
the body and often felt so real that E.P. had difficulties in
distinguishing them from the real limbs. Seeing the real
limb removed the ghost, as did strong tactile stimulation.
Interestingly, the position of the ghost hand copied, in a
perseverative manner, the previous position of the left real
hand: for example, if the left hand had previously been on

the arm rest of a chair but was now on a table, the (proprioceptically) appearing ghost hand was on the arm rest. When
the left hand was moved back to the arm rest, the ghost limb
disappeared but reappeared soon, this time to the table. The
time delay from the new position of the left arm to the
appearance of the ghost limb was measured several times
and was typically 0.5–1 min. If E.P. stayed immobile, the
ghost percept could last for tens of minutes.
The ghost limbs were not under voluntary control but
frequently followed movements of the right limbs. For
example, when E.P. moved one finger of the right hand,
she felt like the corresponding finger would have moved
in the ghost hand at the same time. However, the general
position of the ghost was determined by the previous position of the left hand.
E.P. occasionally experienced splitting of her body: when
she was rising from a bench, she felt that only the right half
started walking, whereas the left side remained on the
bench. Repetitive and monotonical movements triggered
this percept whereas dancing, for example, prevented its
appearance.
E.P.’s neuromagnetic [8] cortical activity was recorded
with a 122-channel whole-scalp neuromagnetometer [1].
The left and right median nerves were stimulated at the
wrists with 0.3-ms constant-current pulses once every 3 s,
with stimulus intensities exceeding the motor threshold.
About 70 artefact-free single responses were averaged
selectively during periods when E.P. perceived or did not
perceive the extra limb. EP was able to provoke the ghost
hand percept with certain positions of the left extremity and
she reported verbally when the ghost appeared/disappeared.
Fig. 3 shows activation strengths of the left primary and
secondary (SI and SII) cortices after right median nerve
stimuli. Both responses occurred at normal latencies. However, responses of the SII cortex [6], strongest about 80 ms
after the stimulus onset, were dampened by 50% during the
ghost hand percept; the finding was well replicable during
successive measurements, and practically identical results
were obtained five months later. The SI cortex showed a
trend to enhanced responses at 115 ms during the ghost
percept whereas the earlier responses were identical in
both conditions. The responses of the right SII (ipsilateral
to the stimuli) were very small and showed no consistent
differences between the conditions. The right-hemisphere
SI and SII responses were normal to left median nerve stimuli but it was not possible to study the ghost limb effect on
them since the stimuli immediately abolished the ghost.
We consider the modification of the SII responses during
the ghost hand percept as an objective correlate of E.P.’s
sensations, especially since the effect is markedly stronger
than the previously described attentional effects on the SII
activity [5]. SII is considered to be related to e.g. haptic
processing, tactile learning and retention, motor intent,
and also to autotopognostic body scheme [4,7,13]. The SII
cortices of both hemispheres are activated after unilateral
limb stimulation and thus SII may have a role in combining

R. Hari et al. / Neuroscience Letters 240 (1998) 131–134

somatosensory information from the two body sides to
allow interhemispheric unification which was apparently
distorted in E.P due to callosal lesion.
Area 5 in the parietal lobe contains representations of
current limb postures [10], which are constantly updated
on the basis of proprioceptive information, visual input,
and probably also by ‘efferent copies’ from the motionrelated brain areas as the limb is deliberately moved. We
cannot exclude the possibility that the 115-ms response
from the SI cortex, stronger during the presence than the
absence of the ghost, would in fact have received contribution form the nearby area 5.
The anterior supplementary motor area, so-called preSMA (area 6ab), is essential for updating motor plans
[12,14–16] and thus a probable source of efferent copies.
Probably due to pre-SMA lesion, signals about intended
movements of E.P.’s left hand fail to update her limb position representation(s); this deficit is certainly accentuated by
the distorted callosal transfer. As a result, E.P. seems to have

Fig. 2. E.P.’s drawing of herself when she has the ghost limb percept.
The upper ghost limb typically originated in the upper left part of the
torso, and the lower ghost limb at the same place as the left lower
limb. Some objects in one real hand (either left or right) were perceived in all three hands and generated embarassing situations:
while shopping E.P. suddenly felt like having three bags, instead of
one, and was anxious of having accidentally taken bags from other
people. E.P.’s description of this figure (translated from Finnish): ‘this
is the feeling I have when I go with a bag or something in my hand. I
have even had three dogs on a lead, although in reality there was
only one’.

133

Fig. 3. Time courses of activation of the left SI and SII cortices after
right median nerve stimulation when E.P. did not perceive the ghost
limb (solid lines) and when the ghost limb was present (dashed
lines). The coordinate systems of the magnetic resonance images
and neuromagnetic data were aligned by means of a 3-D digitizer
and a head position indicator system, and the sources of evoked
signals were modelled as current dipoles [8]. The same measurements were repeated once, shown with the superimposed traces.
The circles on the MRI slice illustrate the dipole locations in the left
SI and SII cortices.

at least two representations of the left arm (leg), which do
not always coincide and lead to the percept of the additional
limb at the previous position of the real left limb. Somatosensory and visual signals still correctly update the representation and cause the ghost to disappear. Signals about the
intended right hand movements seem to incorrectly update
(one of) the left hand representation(s), thereby giving rise
to a moving ghost. The necessity of the efferent copy for
proper updating of the position representation is supported
by the finding that the ghost hand imitates only active, not
passive movements of the right hand.
We hypothesize that due to pre-SMA lesion, E.P.’s representations of limb postures are not properly updated and,
due to callosal disconnection, her hemispheres support different body images. Consequently E.P.’s bodily awareness
is fragmented so that one part relies on current proprioceptive information and another on the previous data which
were relevant tens of seconds earlier. A very similar phenomenon is palinopsia (‘seeing again’) which refers to a
visual perseveration with persistence or reappearance of a
visual image when the original stimulus is no longer present
[11]. Palinopsia has been reported mainly after right-hemisphere lesions. Both palinopsia and the ghost hand percept
of E.P. could be explained by a lack of active suppression of
neural activity.
It thus seems evident that simultaneous mental contents
of body scheme may be based on neural activity which has
emerged at considerably different times and that conscious
body image may be based on competing activity of more
than one brain region. These results are in line with modularity of human bodily awareness. The modules act in concert during normal circumstances but may become
independent, and result in fragmentation of bodily awareness, after distorted updating of cerebral representations.

134

R. Hari et al. / Neuroscience Letters 240 (1998) 131–134

This study was supported by the Academy of Finland.
We thank C. Frith, S. Salenius, G. Curio, O.V. Lounasmaa,
R. Paetau, S. Vanni, and V. Virsu for discussions and
comments, and M. Illman for help with MEG measurements.
[1] Ahonen, A.I., Hämäläinen, M.S., Kajola, M.J., Knuutila, J.E.T.,
Laine, P.P., Lounasmaa, O.V., Parkkonen, L.T., Simola, J.T.
and Tesche, C.D., 122-channel SQUID instrument for investigating the magnetic signals from the human brain, Physica
Scripta, T49 (1993) 198–205.
[2] Banks, G., Short, P., Martinez, J., Latchaw, R., Ratcliff, G. and
Boller, F., The alien hand syndrome. Clinical and postmortem
findings, Arch. Neurol., 46 (1989) 456–459.
[3] Goldberg, G., Mayer, N. and Toglia, J., Medial frontal cortex
infarction and the alien hand sign, Arch. Neurol., 38 (1981)
683–686.
[4] Grafton, S., Fagg, A., Woods, R. and Arbib, M., Functional
anatomy of pointing and grasping in humans, Cerebral
Cortex, 6 (1996) 226–237.
[5] Hari, R., Hämäläinen, H., Tiihonen, J., Kekoni, J., Sams, M. and
Hämäläinen, M., Separate finger representations at the human
second somatosensory cortex, Neuroscience, 37 (1990) 245–
249.
[6] Hari, R., Karhu, J., Hämäläinen, M., Knuutila, J., Salonen, O.,
Sams, M. and Vilkman, V., Functional organization of the
human first and second somatosensory cortices: a neuromagnetic study, Eur. J. Neurosci., 5 (1993) 724–734.
[7] Hyvärinen, J., The Parietal Cortex of Monkey and Man,
Springer-Verlag, Berlin, 1982.
[8] Hämäläinen, M., Hari, R., Ilmoniemi, R., Knuutila, J. and
Lounasmaa, O.V., Magnetoencephalography – theory, instru-

mentation, and applications to noninvasive studies of the working human brain, Rev. Mod. Physics, 65 (1993) 413–497.
[9] Hänninen, R. and Mäkinen, T., Failure of bodily metacontrol:
three hands and three legs or only half of body. A longitudinal
study, 8th World Congress of IOP, Tampere, Finland, 1996,
Volume of Abstracts, p. 131.
[10] Kalaska, J., Parietal cortex area 5 and visuomotor behavior,
Can. J. Physiol. Pharmacol., 74 (1996) 483–498.
[11] Kölmel, H.W., Visual illusions and hallucinations. In C. Kennard
(Ed.), Visual Perceptual Defects, Baillière’s Clinical Neurology,
Baillière Tindal, London, 1993, pp. 243–264.
[12] Marsden, C., Deecke, L., Freund, H.-J., Hallet, M., Passingham, R., Shibasaki, H., Tanji, J. and Wiesendanger, M., The
functions of the supplementary motor area: summary of a workshop. In H. Lüders (Ed.), Supplementary Sensorimotor Area,
Advances in Neurology, Lippincott–Raven, Philadelphia, PA,
1996, pp. 477–488.
[13] Penfield, W. and Jasper, H., Epilepsy and the Functional Anatomy of the Human Brain, Little, Brown and Company, Boston,
MA, 1954.
[14] Rizzolatti, G., Luppino, G. and Matelli, M., The classic supplementary motor area is formed by two independent areas. In H.
Lüders (Ed.), Supplementary Sensorimotor Area, Advances in
Neurology, Lippincott–Raven, Philadelphia, PA, 1996, pp. 45–
56.
[15] Shima, K., Mushiake, H., Saito, N. and Tanji, J., Role for cells in
the presupplementary motor area in updating motor plans,
Proc. Natl. Acad. Sci. USA, 93 (1996) 8694–8698.
[16] Zilles, K., Schlaug, G., Matelli, M., Luppino, G., Schleicher, A.,
Qü, M., Dabringhaus, A., Seitz, R. and Roland, P., Mapping of
human and macaque sensorimotor areas by integrating architectonic, transmitter receptor. MRI and PET data, J. Anat., 187
(1995) 515–537.