Increased Cortical Representation
of
the
Fingers of the Left Hand
in
String Players
Thomas Elbert, Christo Pantev, Christian Wienbruch,
Brigitte Rockstroh, Edward Taub
Magnetic source imaging revealed that the cortical representation of the digits of the left
hand of string players was larger than that
in
controls. The effect was smallest for the left
thumb, and no such differences were observed for the representations of the right hand
digits. The amount of cortical reorganization
in
the representation of the fingering digits
was correlated with the age at which the person had begun to play. These results suggest
that the representation of different parts of the
body
in
the primary somatosensory cortex
of humans
depends
on use and changes
to
conform
to
the current needs and experiences
of the individual.
Evidence
has
accumulated
over
the
past
rwo
decades
that
indicares
that
alterations
in
afferent
input
can
induce
plastic
reorga-
nizational
ch<mges
within
the
adulr
mam-
malian
cenrral
nervous
system
(/).
Ch<lnges
in
the
relation
between
peripheral
sensory
fields
and
their
centra
I
representations
ha
ve
been
observed
for
the
somatosensory
(2),
visual
(I,
3,
4),
and
auditory
systems
(5),
and
comparable
changes
also
have
been
found for
motor
systems
(6).
In
many
of
these
experimeonts,
the
removal
of
afferent
input
from
,1
cortical
region
resulted
in
an
"invasion"
by a
neighburing
area
whose
innervation.
remained
intact.
For
example,
the
cortical
region
representing
a
digit
be-
fore
amputation
in owl
monkeys
could
he
activated
after
amputation
by
tactile
stim-
ulation
of
an
intact
adjacent
finger
(7).
The
changes
noted
were of
the
order
of
a few
millimeters.
More
extensive
plastic
changes
have
recently
been
observed
after
the
abo-
lition
of
input
from larger
portions
of
the
body-for
example,
with
somCltosensory
deafferentation
of
an
entire
forelimb
in
ma-
caque
monkeys
(8)
and
upper
extremity
amputation
in
humans
(9-11).
In
addition,
it
has
heen
shown
in
studies
with
owl
monkeys
that
3
prolonged
increase
of
tactile
stimulation
to
the
distal
pad
of
one
or
two
phalanges
results
in
a
greatly
increased
cortical
representation
speciflc
to
that
portion
of
the
fingers
(12,
13) Evi-
dence
has
also
been
reported
that
suggests
an
increased
cortical
representation
of
the
index
flnger
used in
reading
by
blind
Braille
readers
(14).
Violinists
and
other
string
players pro-
vide
a
good
model
for
the
study
of
the
effects
of
differential
afferent
input
to
the
two
sides
of
the
brain
in
humans.
During
their
pnlCtice
or
performance,
the
second
to
the
fifth digits
(D2
to
05)
of
the
left
hand
are
continuously
engaged
in
fingering
the
strings, a task
that
involves
considerable
manual
dexterity
and
eonhanced
sensory
stimulation.
At
theo
same
time,
theo
thumb
grasps
the
neck
of
the
instrument
and,
al-
though
not
as
active
8S
the
fingers, engages
in
relatively
frequent
small
shifts
of
position
and
pressure.
The
right
h.and,
which
manip-
ulates
th.eo
bow,
participates
in a task
involv-
ing
much
less
individual
finger
movement
and
fluctuation
in
tactile
and
pressure in-
put.
Here,
we
present
data
from
magnetic
source
imaging
that
indicates
that
the
cereo-
bral
cortices
of
string
players
3reo
different
from
the
cortices
of
controls
in
that
the
representation
of
the
digits
of
the
left
hand
is
substantially
enlarged
in
the
cortices
of
string
players.
Nine
musicians
(six violinists,
two
cel-
lists,
and
one
guitarist)
who
had
played
their
instruments
for a
mean
pcoriod
of
11.7
years
(range,
7
to
17 years) served as sub-
jects
for
our
study.
Six
nonmusicians
served
as
controls
(/5).
The
mean
age for
bm
l
,
groups was 24
:!:
3
yecltS.
Before
our
im',
tig,ltion,
the
musicians
kept
a diary for 1
week,
recording
the
amount
of
time
prac-
ticed
per
day
(mean
9.8
:!:
8.4
hours
per
week),
and
had
estimated
the
amount
of
ti
me
spent
practicing
during
the
preovious
month
and
year
(l0.8
:!:
8.8
hours
per
week)
During
theo
experimental
session, so-
matosensory
stimulation
was
delivered
to
the
first
digit
and,
in
separate
runs, to
the
fifrh digit
of
either
hand.
Stimulation
con-
slsted
of
light
superficial
pressure
applied
by
means
of
a
pneumatic
stimulator
with
the
use
of
standard,
nonpainful
stimulation
in-
tensity
(9,
16, 17)
The
daw
(Fig 1)
indi-
cate
that
the
center
of
cortical
rcsponsivity
for
tactile
stimulation
of
the
digits
of
the
left
hand
was
shifted
in
musicians
as
com-
pared
to
that
in
controls,
while
at
the
same
time
the
strength
of
response
increased.
The
topographic
shift
was
toward
the
mi,J
sagittal
plane,
which,
along
the
surface
th~
postcentral
gyrus,
is
toward
the
region
30
B
25
••
..
•
String players
20
15
•
..
10 •
@
I
5
O+-~'--'--",-"'-",-"'-r-..---,-
o 0.5 1 1.5 2 2.5
Distance,
01
to
05,
left
hand (cm)
Controls
.~
.
~@
•
•
@
,.
String players
E 3 C
~
-g
2.5
'"
.J::
:g,
2
.;:
~~
1.5
.8
Ei
2f
0.5
c:
~
'"
is
E
~
.s
.J::
c,
c:
e
U;
III
(5
C.
1J
'"
Cl Controls
O-j-.,...~..,--~..........,~~...".-~~
o 5 10 15
20
Age at
inception
of
musical
practice
Fig.
1.
(A)
Equivalent current dipoles elicited
by
stimulation
of
the thumb (01)
and
fifth
finger (05)
of
the
left
hand are superimposed onto an
MRI
(magnetic
resonance imaging) reconstruction
of
the cerebral cortex
of
a control, who
was selected
to
provide anatomical landmarks
for
the interpretation
of
the
MEG-based localization. The arrows represent the location and orientation
of
the
ECO
vector
for
each
of
the two digits averaged
across
musicians (black)
and controls
(yellow).
The length
of
the arrows represents the mean magni-
tude
of
the dipole moment
for
the two digits
in
each
group. The average
locations
of
05
and
01
are shifted medially
for
the string players compared
to
controls; the
shift
is
larger
for
05
than
for
01.
The dipole moment
is
also larger
for
the musicians'
05,
as
indicated by the greater magnitude of the black
arrow.
(8) The magnitude
of
the dipole moment
as
a function of the
age
of
inception
of
musical practice; string players are indicated
by
filled
circles,
control subjects by hatched circles. Note the larger dipole moment
for
indi-
viduals beginning musical practice before the
age
of
12.
(C)
Scatterplot
of
the
Euclidean distances
(in
centimeters) between the cortical representations
of
01 and 05. This distance
for
the musicians'
left
hands was greater than that
in
controls, but this difference
is
not statistically significant.
SCIENCE • VOL.
270
•
13
OCTOBER J
995
305
Konstanzer Online-Publikations-System (KOPS)
URL: http://www.ub.uni-konstanz.de/kops/volltexte/2008/6428/
URN: http://nbn-resolving.de/urn:nbn:de:bsz:352-opus-64286
First publ. in: Science 270 (1995), pp. 305-307
11
of
the
cortex
that
represents
the
palm
of
the
hand
(18).
For
OS
(little
finger)
of
the
left
hand,
the
shift
was 0.7
cm
(t = 3.6, P <
0.01, degrees
of
freedom = 13); for
01
(thumb),
the
shift was
0.5
cm
(t = 3.3, P <
0.0l).
This
shift
of
OS
was significantly
greater
than
the
shift
of
01.
Correspondingly,
the
analysis
of
vari-
ance
(ANOVA),
which
included
data
from
both
hemispheres
(18),
showed
an
interac-
tion
of
group (a factor
in
ANOVA)
and
digit
[F(l,13)
= 4.78, P < 0.05].
This
effect
was
dependent
on
the
side
on
which
stim-
ulation
was given:
There
were
no
signifi-
cant
shifts for
the
digits
of
the
right
hand
of
the
musicians
compared
to
those
of
the
contr,)ls.
The
dipole
moment,
which
is
pre-
sumed
to
be
an
index
of
total
neuronal
~ctivity,
also increased for
the
stimulation
o'-..J
the
digits
of
the
left
hand
of
musicians
compared
to
the
left
hand
of
controls
[ANOV
A
interaction
of
group
and
side
of
stimulation,
F(l,
13) = 5.54, P < 0.05j.
The
increase was larger for
OS
(t = 5.4, P <
0.0l)
than
for
Dl
[t
= 2.0, P < 0.1;
ANOV
A
interaction
of
group
and
digit,
F(l,13)
=
4.81,
P <
0.05]
(19).
There
was a
correlation
between
the
age
at
which
the
string
players began studying
their
instruments
and
the
magnitude
of
the
change
in
the
dipole
moment
of
05
com-
pared
to
that
in
controls
(r = 0.79, r =
0.0l)
(Fig.
lB).
The
relation
between
the
amount
of
practice
and
the
cortical
mea-
sures was
not
significant.
The
increase
of
the
dipole
moment
of
both
01
and
05
in
the
string players indicates
that
an
extend-
ed
cortical
network
responds
to
tactile stim-
ulation.
If
the
active
area expands,
the
point
location
determined
by
an
equivalent
"Dole
model
will shift inward.
Given
the
~erical
geometry
of
the
head,
measure-
ments
of
the
01
and
05
representations
would
approach
each
other
if
the
center
of
activity
remained
unchanged.
Conversely,
an
increase in area of
representation
would
produce
opposing
effects
on
the
measured
cortical
distance
between
01
and
D5,
which
could
explain
the
absence
of
a sig-
nificant
group difference
in
the
measured
distance
between
cortical
representations
of
01
and
05
(Fig. 1
C).
Further
experiments
should
attempt
tl)
model
the
size
of
the
activated
area
to
resolve this
question.
The
significant shift in medial
direction
of
the
cortical
representation
of
the
fingers
of
the
left
hand
in
string
players
and
the
increase in
the
corresponding
dipole mo-
ments
suggest
that
the
cortical
territory oc-
T.
Elbert and
B.
Rockstroh, Department of Psychology,
University of Konstanz, 0-78434 Konstanz, Germany.
E-mail: Thomas.Elbert@Unl-Konstanz.de
C.
Pantev and
C.
Wienbruch, Biomagnetism Center, Uni-
versity of Munster,
0-48129
Munster, Germany.
E.
Taub, Department of Psychology, University of Ala-
bama
at
Birmingham, Birmingham, AL 35294, USA.
306
cupied
by
the
representation
of
the
digits
increased
in
string players
as
compared
with
that
in controls.
Two
alternate
interpreta-
tions
of
our
data
sh'luld
be
taken
into
ac-
count.
First, it
could
be
argued
that
individ-
uals
with
a genetically
determined
large
representatil11l
of
the
left
hand
digits make
superior
string
players
and
therefore
are
more
likely to
continue
with
musical
train-
ing
once
they
have
begun. However,
in
research
with
animals,
use-dependent
en-
largements
of
portions
of
the
somatosensory
map
in
cortical
area
3b
have
been
clearly
demonstrated
under
conditions
of
increased
use generally similar
to
those
in
this
study
(12, 13).
In
either
case,
the
relatively larger
representations
of
individually
important
digits could
have
the
role
of
enhancing
the
particular
needs
of
a string player
in
an
adaptive
manner.
A second
alternative
explanation
of
our
results
is
that
they are a consequence
of
a
shift
in
cortical responsivity
combined
with
an
intensification
of
the
response. However,
we
think
a more plausible
explanation
is
that
the
c,)[tical territory
of
the
left-hand digits
has expanded.
This
is
more plausible because
(i)
there
is
a
correlation
between
amount
of
cortical reorganization
and
age [or stage
of
central
nervous system
(CNS)
maturation]
at
which
musical practice
began
and
(ii)
the
equivalent
current
dipole
(ECD)
shift fol-
lows
the
one
direction
that
is
consistent
with
the
expansion
interpretation.
Related work (10) has shown
that
there
is
a strong correlation
in
humans with upper
extremities amputated between
extent
of
cor-
tical rellrganization
and
amount
of
phantom
limb
pain
experienced.
Although
phantom
limb
pain
is
a maladaptive result of nervous
system injury, our results demonstrate the
functional relevance
of
cortical reorganiza-
tion, similar
to
results reported
in
the
context
of
the visual sysrem (3)
and
auditory system
(12,
20,
21
).
One
may speculate
that
one
role
of cortical reorganization might be
to
contrib-
ute to
the
functional recovery
of
organisms
after
CNS
damage, possibly in terms
of
recov-
ery from
CNS
shock.
The
evolutionary ad-
vantage
of
this mechanism
is
brought into
question, however, by
the
fact
that
the
process
of recovery
is
usually slow
and
thus would
not
permit
an
organism
that
was seriously im-
paired to survive long enough
to
engage in
sllccessful reproduction
and
transmission
of
this capacity (22).
However,
in accord
with
the
results
of
Merzenich
and
co-workers (7),
continuous
plastic reorganization
of
cortical
space
that
permits
rapid
reallocation
of
available
CNS
circuitry would
confer
an
obvious
practical
advantage.
The
possible
contribution
of
cortical
reorganization
to
recovery
of
func-
tkm
after
CNS
injury
might
thus
be
an
adventitious
result
that
"piggy
backed"
onto
a
mechanism,
which
permitted
the
SCIENCE •
VOL.
270
•
13
OcrOBER
1995
much
more critically
important
plastic pro-
cesses associated
with
learning, physical
growth
during
maturation,
and
adjustment
to
current
environmental
demalllk
The
role
of
this
mechanism
in
the
recovery
of
function
would
become
impl)rtant
only
when
protecting
an
individual
with
CNS
damage would artificially
prolong
survival
long
enough
for
extensive
cortical
reorga-
nization
to
work.
REFERENCES AND NOTES
1.
J.
H.
Kaas, Annu.
Rev.
Neurosci.
14,137
(1991).
2.
A.
Mogilner
et
al., Proc. Nat!. Acad. Sci.
USA
90,
3593 (1993); M.
M.
Merzenich
et
al
.. Neuroscience
8,33
(1983);
S.
A.
Clark.
T.
Allard,
W.
M.
Jenkins,
M. M.
Merzenich, Nature 332,
444
(1988); J.
Kalaska and
B.
Pomeranz, J. Neurophysiol. 42,
618
(1979);
D. D.
Rasmusson, J. Comp. Neural.
205,313
(1982);
A.
M.
Kellohan and G.
S.
Doetsch,
Somatosens
Res.
2,49
(1984);
J.
T.
Wall and
C.
G.
Cusick,
J.
Neurosci.
4,
1499 (1984); J.
T.
Wall et al.,
Ibid.
6,218
(1986).
3.
C.
Darian-Smith and
C.
O.
Gilbert, Nature 368, 737
(1994).
4.
C.
D.
Gilbert and
T.
N.
Wiesel, ibid. 356. 150 (1992);
F.
Rosier,
B.
Rbder,
M.
Heil,
E.
Henninghausen, Cog-
nit. Brain
Res.
1, 145 (1993).
5.
G.
H.
Recanzone,
C.
E.
Schreiner,
M. M.
Merzenich,
J. Neurosci
13,87
(1993);
D.
Robertson and
O.
R.
F.
Irvine, J. Comp. Neurol. 282, 456 (1989);
D.
R.
F.
Irvine,
R.
Rajan,
L.
Z.
Wize,
P.
Heil, Soc. Neurosci.
Abstr.
17, 1485 (1991).
6.
L.
G.
Cohen,
S.
Bandinelli,
T.
W. Frindley,
M.
Hallet,
Brain 114, 615 (1991);
J.
N.
Sanes,
S.
Sun
er,
J.
F.
Lando,
J.
P Donaghue, Proc. Natl. Acad.
Sci.
USA
85, 2003 (1988).
7.
M. M.
Merzenich et al., J. Comp. Neurol.
224,591
(1984).
8
T.
P.
Pons et al., Science 252, 1857 (1991).
9.
T.
Elbert
et
al., Neurarepol1
5;
2593 (1994).
10.
H.
Flor
et
al., Nature 375, 482 (1995)
11.
T.
T.
Yang
et
aI., Neurorepol15,
701
(1994).
12. G.
H.
Recanzone,
M.
M.
Merzenich,
W.
M.
Jenkins,
A.
Grajski,
H.
R.
Dinse, J. Neurophysiol. 67,
1031
(1992)
13.
W.
M.
Jenkins,
M.
M.
Merzenich,
M.
T.
Ochs,
T.
Allard,
E.
Guic-Robles, ibid. 63, 82 (1990).
14.
A.
Pascual-Leone and
F.
Torres, Brain 116, 39
(1993).
15. Controls were either students or university employ-
ees who worked
in
an
academic environment and
had no experience playing string instruments. They
were selected to be comparable
to
the experimental
group with respect
to
age, social status, and gender.
After giving informed consent, experimental partici-
pants were briefly interviewed concerning their mu-
sical activities and their history of practicing. Experi-
mental participants were also interviewed
to
elicit
information regarding their usual manual activities:
None of the controls engaged
in
manual tasks to
an
unusual extent oUler than typing on computer key-
boards. Each of them used both hands for word
processing orothercomputeroperations. Two
of
the
rnusicians were found to have a tendency toward
ambidexterity, but
all
other test participants were
right-handed (Edinburgh handedness questionnaire)
[R.
C.
Oldfield. Neuropsychologia
9,
97
(1971)J.
16.
T.
T.
Yang,
C. C.
Gallen,
B.
J. Schwartz,
F.
E.
Bloorn,
Proc. Natl. Acad. Sci.
USA.
90,3089
(1993).
17. A BTI-Magnes system was used for MEG (magneto-
encephalographic) recordings. The sensor array was
positioned over one of the two hemispheres (C3 or
C4)
in
a fixed irregular order. At each site, 1000
stimuli were delivered at
an
average rate of 0.5 Hz
(the interval between stimulus onsets was 500
:'::
50
rns).
Stimulation site sequence vaned according to a
fixed irregular ordm across test participants. Within
the range of 30
to
75 ms, a first major peak was
identified
in
each of the evoked wave forms. The
mean latencies were 46.2:':: 7.9 ms for left
01,
and
48.1
:'::
7.2 ms for right
01;
for left 05, the mean
latency was 52.2
:+:
8.3 ms, and for right
05
52.8
:+:
7.5 ms. The difference between musicians and con-
trols for latency was not significant. For each evoked
magnetic field, a single
EGO
model (best fitting local
sphere) was fitted and the medians of the dipole
moment and
the
dipole location were computed
from a selection of points within a 20-ms time seg-
ment
(11
sampling points) around the maximal rms
(root mean square across the 37 channels) within the
range of 30
to
75 ms.
POints
were selected if they
met the
follOWing
requirements:
(i)
rms indicating a
signal-to-noise ratio
>3;
(ii)
a goodness of fit of the
EGO
model to the measured field
>0.95;
and
(Iii)
a
minimal confidence volume of the
EGO
location
<300
mm'J,
18. For
all
cortical measures,
an
ANOVA with the be-
tween-subJect factor group (musicians versus con-
trols) and the within-subject factors digit
(01
versus
05) and side of stimulation (left versus right) was
computed first. ANOVAs for subsets of the data
or
t
tests were used to resolve interactions.
19.
Given a constant direction of the equivalent current
dipole, the dipole moment indicates the total
strength of cortical
polarization-that
is,
tile number
of neurons involved during a cortical response.
If
this
number increases,
the
dipole moment also increas-
es.
Any active focal area can be modeled by
an
equivalent current dipole. Each dendritic current flow
contributes
to
this dipole moment according to the
formula
dipole moment
= (conductivity)
x (cross section
of
the
dendrite)
x (potential difference along
the
dendrite)
S.
J. Williamson and
L.
Kaufman
[in
Auditory
Evoked
Magnetic
Fields
and
Electric Potentials,
F.
Grandori,
M. Hoke,
G.
L.
Romani, Eds. (Karger, Basel, 1990),
pp.
1-39]
assume the diameter of
an
apical dendrite
to
be 4
fCm,
the intracellular conductivity to be about
0.25
S/m,
and the potential difference
to
be about 10
mV.
With the use of these assumptions, about
30,000 dendrites would be necessary to produce a
dipole moment of 10 nA-m
If
conductivity and po-
tential difference are not different
in
musicians and
controls, the magnification of the dipole moment
in
response
to
finger stimulation of the left hand
in
mu-
sicians can be explained
if
approximately twice as
many cells were activated
in
musicians than were
activated in the controls.
20. G.
H.
Recanzone, M. M. Merzenich, J. Schreiner,
J.
Neurophysiol. 67, 1071 (1992)
21.
N.
M.
Welnberger
et
al.,
Concepts
Neurosci. 1,
91
(1990)
22. We are indebted
to
T.
Pons for this observation.
23. We appreciate the assistance of
S.
Hampson,
B.
Lutkenh6ner, and
O.
Steinstrater. Supported by the
Oeutsche Forschungsgemeinscllaft.
25 May 1995; accepted 13 September 1995
r
..
",
SCIENCE
VOL.
270
13
OCTOBER 1995
,/
307