Ad ances in Elec ical and Elec onic Enginee ing
310
THE POTENCY OF OPTICAL AND AUGMENTED REALITY MIRROR BOXES IN
AMPUTEES AND PEOPLE WITH INTACT LIMBS
K. O’Neill
1,2
, A. de Pao
1,2
1 School o Elec ical, Elec onic and Mechanical Enginee ing, Na ional Uni e si y o I eland, Dublin (UCD),Bel ield,
Dublin 4, I eland
2 Rehabili a ion Enginee ing, Na ional Rehabili a ion Hospi al, Dún Laoghai e, Co. Dublin
e-mail:ann ao[email p o ec ed]e
Summa y. Phan om limb pain is he dis essing p oblem expe ienced by many ampu ees, de ined as a pain ul
sensa ion pe cei ed in he a ea o he missing body pa . Phan om limb pain can be e y se e e and disabling. I
con inues o be expe ienced by wo hi ds o ampu ees, eigh yea s pos -ampu a ion. Augmen ed eali y has he
abili y o change a pe son’s senso y expe ience. Mo e applica ions o his echnology a e g adually being
u ilised o he apeu ic pu poses as augmen ed en i onmen s can be used bo h o dis ac he a en ion o pa ien s
om exc ucia ingly pain ul expe iences and o p omo e co ical e-mapping a he si e om whe e he pain
a ises. Using Augmen ed Reali y, an en i onmen has been c ea ed whe e uppe limb ampu ees can bo h iew
and con ol mo ion o hei phan om limb o help alle ia e phan om limb pain.
1. INTRODUCTION
The sensa ion o a phan om limb can be desc ibed as
he eeling o he p esence o an ampu a ed
ex emi y as s ill p esen . This phenomenon o a
phan om limb has been ecognised o well o e a
cen u y [1,2]. Almos e e yone who has a limb
emo ed will expe ience a phan om limb, wi h mos
ampu ees ha ing epo ed expe iencing a phan om
immedia ely a e he loss o a limb [3].
Immedia ely a e ampu a ion, he phan om limb
usually esembles he p e-ampu a ion limb in shape,
leng h and olume. The sensa ion can be desc ibed
as e y i id and o en includes eelings o pos u e
and mo emen . In many cases, he phan om limb is
p esen ini ially o only a ew days o weeks, be o e
i g adually ades away om consciousness. Once
he phan om ades away i gene ally does so
comple ely, ye some ampu ees can ecall he
phan om by ubbing he s ump o by in ense
concen a ion [2]. Howe e , he phan om limb can
also pe sis o yea s and may ne e ade away.
30% o pa ien s s ill ha e hei phan om limb
decades la e [4] and case epo s exis ha de ail
phan om limbs ha ha e pe sis ed o 44 yea s [5].
Phan om limb pain occu s when a pe son eels pain
ha is a ibu ed o he a ea o a limb ha has been
emo ed. Phan om limb pain can be se e e and
disabling and con inues o be expe ienced by wo
hi ds o ampu ees, eigh yea s pos ampu a ion [6].
As many as 70% o phan oms s ill emain pain ul 25
yea s a e he loss o he limb [7]. Phan om limb
pain occu s in a iable o ms. In some ampu ees he
pain is con inuous bu a ying in in ensi y, while
o he s expe ience in e mi en pain o a high
in ensi y [8]. Phan oms end o be mo e i id, and
pe sis longe , a e auma ic limb loss, o ollowing
ampu a ion o a p e-exis ing pain ul limb
pa hology, han a e a planned su gical ampu a ion
o a non-pain ul limb.
I is o en epo ed ha a phan om limb will occupy
a habi ual pos u e, bu spon aneous changes in
posi ion a e also common. The phan om may also
assume, ei he empo a ily o pe manen ly, an
awkwa d and pain ul pos u e. Phan oms a e mo e
i id and pe sis longe a e auma ic limb loss, o
ollowing ampu a ion o p e-exis ing pain ul limb
pa hology, han a e a planned su gical ampu a ion
o a non-pain ul limb. This may be due o he
g ea e a en ion paid o he mu ila ed o pain ul
limb be o e i is los . Memo ies o he limb’s
pos u e and o m p io o ampu a ion o en su i e
in he phan om [9]. In addi ion, a e ampu a ion o
a de o med limb, he de o mi y is o en ca ied o e
in o he phan om [4,10].
Telescoping is a phenomenon ha is common in
app ox. 50% o cases, especially in uppe -limb
ampu ees. This is whe e he dis al pa o he limb is
g adually el o app oach he esidual limb. In he
case o an uppe limb ampu ee, he phan om a m
ge s sho e o e ime un il he pa ien is le wi h
jus a phan om hand a ached o hei s ump [11,12].
2. CORTICAL BRAIN REMAPPING AFTER
AMPUTATION
The s udy o phan om limbs also p o ides an
oppo uni y o unde s and exac ly how he b ain
cons uc s a body image, and how his image is
con inuously upda ed in esponse o changing
senso y inpu s. A comple e soma o opic map o he
body su ace exis s in he soma osenso y co ex o
p ima es [13,14], including humans [15] (See Figu e
1). The soma osenso y co ex localises he sou ce o
senso y inpu and pe cei es he le el o in ensi y o
he s imulus. I is capable o spa ial disc imina ion,
so i can disce n shapes o objec s being held o
dis inguish sub le di e ences in simila objec s ha
come in o con ac wi h he skin.
The po ency o op ical and augmen ed eali y mi o boxes…
311
Fig. 1: Pen ield Homunculus shows whe e he si e o
each limb is loca ed on he soma osenso y co ex o he
b ain
I has been shown ha co ical eo ganisa ion occu s
a e ampu a ion [16] and ha he e is a high posi i e
co ela ion be ween he magni ude o phan om limb
pain and he amoun o co ical eo ganisa ion [17].
S udies o a m ampu a ion using
magne oencephalog aphy ha e shown ha b ain
a eas, which o dina ily ep esen he hand, we e
ac i a ed when ei he he lowe ace o uppe a m
was ouched [18,19]. S udies ha e shown ha when
he ace o an ampu ee was s imula ed by ouch,
senso y in o ma ion wen no only o he acial a ea
o he co ex, bu i also in aded he si e ha , p io o
ampu a ion, had ep esen ed he hand. The Pen ield
Homunculus can be eo ganised o e a dis ance o
2-3cm in he adul b ain, ollowing loss o a limb
[20,21,22]. In his way, he b ain ci cui y can be
al e ed.
3. MIRROR BOX ILLUSION
V.S. Ramachand an de eloped a echnique o ea
phan om limb pain in uppe limb ampu ees by using
an o dina y mi o [23]. The box is made by placing
a e ical mi o inside a ca dboa d box wi h he oo
o he box emo ed. The on o he box has wo
holes in i , h ough which he subjec inse s his/he
good a m and he phan om a m. (See Figu e 2)
When iewed om sligh ly o -cen e, he e lec ion
o hei “good” a m ga e he imp ession ha he
subjec had wo in ac a ms. By sending mo o
commands o he in ac a m o make symme ic
mo emen s, as i conduc ing an o ches a, i gi es
he imp ession ha he phan om a m has esu ec ed
and he subjec ecei es posi i e isual eedback
in o ming he b ain ha he phan om a m is mo ing
in synch ony. This p ocess has been dubbed he
Mi o Box Illusion and is he main basis o his
s udy.
Fig. 2: Demons a ing he Mi o Box Illusion
The e is g ea a iabili y in he expe ienced
au hen ici y o he mi o box illusion and i s abili y
o alle ia e phan om limb pain. The e ec o his
posi i e isual eedback o he ampu ee can be e y
he apeu ic. In some cases i was possible o
eposi ion phan om limbs ha we e pe cei ed o be
held in pain ul o awkwa d posi ions, in o non-
pain ul pos u es, gi ing empo a y elie om
phan om limb pain. Whils his echnique had qui e
d ama ic he apeu ic alue o some people, i was
only mode a ely e ec i e, o comple ely ine ec i e
o o he s [23].
4. AUGMENTED REALITY MIRROR BOX
The mi o box illusion had qui e d ama ic
he apeu ic alue o some people, bu i was only
mode a ely e ec i e, o comple ely ine ec i e o
o he s. Fo hose ha ound i o be bene icial, i
was possibly because he illuso y phan om p o ides
eedback ha is consis en wi h ac ual phan om
expe iences. Howe e ampu ees’ pe cep ions o
hei phan om limbs o en di e g ea ly om hei
o iginal limbs [24]. The phan om limb can be
sho e , o longe , a y in hickness, ha e gaps o be
con inuous, in compa ison o he o iginal limb. I is
no unhea d o o a phan om limb o be
disconnec ed om he ampu ee’s s ump (See Figu e
3). This is one eason why he mi o box illusion
does no ha e any he apeu ic alue in some cases,
as i is no uncommon o he e lec ed image o bea
no esemblance o he phan om limb pe cei ed by
he ampu ee. These i egula ly shaped phan oms
canno be iewed in he mi o , as i necessa ily
e lec s he image o he in ac a m.
Fig. 3: The Phan om limb o his Ampu ee is pe cei ed o
esemble a hand disjoin ed om he elbow s ump [24]
Ad ances in Elec ical and Elec onic Enginee ing
312
We ha e c ea ed an en i onmen using Augmen ed
Reali y ha simula es he mi o box illusion. I
consis s o a 3-d g aphical ep esen a ion o an a m
on a la sc een ha is con olled by a wi eless da a
glo e (5DT Da a Glo e 5W, T ansmission
F equency 433.92MHz.). The appea ance o he
a m can be al e ed o esemble he phan om limb o
he use . The da a glo e is wo n on he in ac a m,
whils he phan om appea s on a la compu e
sc een ha akes he place o he mi o . As he
in ac a m mo es, he i ual a m on he sc een
ep esen ing he phan om also mo es in unison.
To c ea e he i ual ep esen a ion o an a m, we
used he 3D G aphics anima ion package 3DS Max
o de o m a single mesh objec . By a p ocess
e e ed o as “Skinning”, we added a bone s uc u e
o a de o mable mesh. When a bone is assigned o a
mesh, i is gi en a se o e ices. These e ices a e
weigh ed o de e mine how much in luence each
bone has on he mesh. We wan ou mesh o bend a
he join s o he a m, bu no in he middle o he
o ea m when he e is mo ion in he bones (See
Figu e 4). Skinning is he mos in ica e pa o he
design p ocess as i dic a es how he mesh will
de o m unde he in luence o each bone.
Fig. 4: We se he in luence o he bones on he e ices in
o de o ge he mesh o de o m p ope ly
The main ad an age o he Vi ual Mi o is ha all
ypes o phan om limbs, such as Figu e 3, can be
g aphically ep esen ed on he sc een. In e ec ,
phan om limbs can be ailo ed o isually esemble
wha he ampu ee pe cei es hem o be. Ano he
ex a bene i ha he Augmen ed Reali y mi o has
o e he gene ic Mi o Box is ha i can also be se
up o gi e mo e gene al mo emen s, o example,
mo ion o he i ual image in he same di ec ion as
he in ac a m, o ha ing a hi d pa y con ol he
mo ion o he image on he sc een.
5. EXPERIMENTAL STUDY
Fou ampu ees and a con ol g oup o 22 able-
bodied people pa icipa ed in he s udy unde
condi ions o in o med consen and in acco dance
wi h e hical p ocedu es.
Ampu ee 1 is a 40 y old male who unde wen igh
side abo e elbow ampu a ion necessi a ed by
os eogenic sa coma. A e ampu a ion, he had he
sensa ion ha his a m was in ac and in i s no mal
posi ion. G adually, o e he pe iod o six o se en
mon hs he expe ienced elescoping o his phan om.
He now pe cei es his phan om hand o be a ached
o he end o he esidual limb, he sec ion o he
phan om ha o iginally joined he phan om hand o
he esidual limb ha ing disappea ed o e ime. The
phan om hand is el o be c amped in a claw-like
posi ion, which al hough discom o ing is no
pain ul.
Ampu ee 2 is a 25 y old male who sus ained an
ampu a ion o he igh o equa e and mino inju y
o his igh leg when his mo o bike collided wi h a
ca whose d i e was in oxica ed. His phan om a m
is posi ioned ac oss his body as hough aised in a
p o ec i e pose. Sensa ion is mos i id in he hand,
which is held in a loosely clenched is wi h he
humb on he ou side. Al hough he does no eel he
o ea m and uppe a m pe se, he is awa e o he
posi ion o he elbow. The phan om does no make
in olun a y mo emen s and he canno olun a ily
gene a e mo emen s in his phan om limb.
Ampu ee 3 is a 64 y old emale who sus ained
inju ies in a ehicle acciden leading o he
subsequen ampu a ion o he igh a m below he
elbow. She expe ienced phan om sensa ion
immedia ely a e ampu a ion su ge y. She is no
o dina ily awa e o sensa ion in he missing o ea m
bu is awa e o he phan om hand. He inge ips a e
el o be ouching he humb, and al hough
o iginally el o be ouching he humb, he li le
inge is now el o be aised. She does no
expe ience he phan om inge s indi idually bu
a he as one uni . When pain ul, he phan om hand
is el as a clenched is wi h he nails digging in o
he palm o he hand.
Ampu ee 4 is a 42 y old emale who unde wen
igh side abo e elbow ampu a ion a e a a ic
acciden . She expe ienced he sensa ion o a
phan om limb immedia ely a e ampu a ion. O e
ime he phan om hand has elescoped, so ha she
now pe cei es i o eside a he end o he esidual
limb. When she expe iences phan om pain, he
phan om hand is el as a clenched is wi h he nails
digging in o he palm o he hand.
Fo he con ol g oup, he subjec s consis ed o 8
males and 14 emales, anging in age om 20 o 29
(mean age = 24.91, s anda d de ia ion = 12.72).
The pa icipan s we e asked o d aw a ca d om an
unbiased deck o de e mine whe he hey would s a
wi h he s anda d mi o o he Augmen ed Reali y
e sion. This is o ensu e ha andomness was
obse ed in he o de asks ha we e pe o med.
(The p ocedu e desc ibed below is based on a
subjec i s pe o ming he asks wi h he s anda d
mi o box, be o e epea ing hem wi h he
Augmen ed Reali y e sion. Fo hose ha selec ed
The po ency o op ical and augmen ed eali y mi o boxes…
313
o pe o m he asks wi h he Augmen ed Reali y
mi o box, he o de was simply e e sed).
Ini ially, each subjec was in oduced o he mi o
box illusion and he/she ied o expe ience he
sensa ion o a phan om limb by opening and closing
hei le hand in on o he mi o , as i hey we e
conduc ing an o ches a and also o lex hei inge s
acco dingly. The pa icipan s we e hen equi ed o
y he mi o box illusion o 3 pe iods o 30
seconds, du ing which hey we e ins uc ed o
epea edly close hei le hand in o a is , pause
b ie ly be o e opening i ou again in on o he
mi o . Then hey held hei hand ou s aigh and
ben he hi d inge inwa ds be o e eleasing i .
This ac ion was eplica ed by a andomly chosen
inge (o he han he hi d digi ). This p ocess was
epea ed o 3 pe iods o 30 seconds du a ion.
Finally, he subjec s we e ins uc ed o place hei
hand in on o he mi o la on a able and o
d um hei inge s hy hmically agains he la
su ace.
The pa icipan s we e hen i ed wi h he le -
handed 5DT da a glo e. They posi ioned hei le
a m in on o he la sc een and hei igh a m
behind i . An image o a igh a m appea ed on he
sc een. The image used he e was a ealis ic 3-d
ep esen a ion o a human a m. The subjec s we e
ins uc ed o keep hei eyes ixed on he image on
he sc een igh h oughou he expe imen . The
same expe imen al p ocedu es as be o e we e
ca ied ou , wi h he la sc een aking he place o
he mi o .
Fig. 5: Subjec using he Augmen ed Reali y Mi o Box,
whils da a is being eco ded
Finally, he subjec s we e asked o epea he
expe imen s again using he Augmen ed Reali y
mi o box. Howe e , his ime a ca oon-like image
o an a m appea ed on sc een, di e en in p opo ion
and shape o hei in ac a m.
We measu ed he EEG signal om he si e on he
co ex co esponding o he igh hand o de e mine
whe he any ac i i y was occu ing in his egion.
The mu hy hm was eco ded o e he senso imo o
co ex o he b ain. The main elec ode placemen s
o his a ea a e along he C line o he 10-20 sys em
o elec ode placemen wi h C3 and C4 being used
o eco ding o changes o mu hy hms associa ed
wi h hand mo emen s [25].
Du ing ini ial mi o box expe imen s i was no ed
ha he igh hand o some o he subjec s (held
behind he sc een/mi o ) mo ed as he image on he
sc een did. We ied o ack his mo ion by using
wo small ligh -sensi i e accele ome e s. These
we e aped o wo sepa a e himbles, which we e
placed on he middle inge ip o each hand while
he expe imen s we e being pe o med. We also
eco ded an EMG signal om he igh a m as in
some cases he mo emen o he igh hand was oo
sligh o be obse ed by accele ome e s.
The EEG, EMG and accele ome e signals we e
only eco ded om he con ol g oup, as he idea
behind hem only e ol ed a e es ing wi h he
g oup o ampu ees. I is hoped o eco d hese
eadings om he nex g oup o ampu ees. These
cap u ed signals we e eco ded using a Biopac
sys em and we e ed in o MATLAB ia a PCI-
6023E da a acquisi ion ca d om Na ional
Ins umen s.
6. RESULTS
Ampu ee 1 ound ha his phan om ini ially became
ense while doing he expe imen s. I p oduced a
pain ul sensa ion in he digi s ha we e mo ing, bu
a elaxed sensa ion in he pa s o his phan om ha
we e a es . The Augmen ed Reali y mi o had he
same e ec o a lesse deg ee and esul ed in a less
in ensi e phan om sensa ion han when he mi o
was used. He epo ed ha he igh sensa ion in his
phan om began o diminish and become milde o e
ime, bu ha his phan om sensa ion became much
s onge in all o his inge s.
P io o es ing Ampu ee 2 was unable o p oduce
olun a y mo emen in he phan om. The p o ision
o eedback, whe he using a s anda d mi o o he
augmen ed eali y box, p o ided him wi h no change
in he posi ion o he phan om and did no al e his
oli ional con ol o his phan om limb.
Fo Ampu ee 3 isual eedback bo h in he s anda d
mi o and augmen ed mi o condi ions in ensi ied
and ex ended phan om expe ience, hough i was
much mo e p onounced when using he augmen ed
en i onmen . When using he s anda d mi o , she
epo ed ha she could i idly eel he palm o he
phan om hand, and ha his sensa ion could be el
igh up h ough he a m. Viewing he augmen ed
eali y phan om image acili a ed mo emen o he
phan om o e inge as an independen digi . Such
mo emen had p e iously been impossible.
Mo emen o he inge s o he phan om hand could
be emo ely gene a ed.
Ampu ee 4 expe ienced he sensa ion o he
phan om hand mo ing using bo h mi o echniques,
bu ound ha i was mo e i id when using he
Augmen ed Reali y mi o . She had ne e be o e
expe ienced olun a y mo ion o he phan om limb.
She was able o p oduce small mo ions o opening
Ad ances in Elec ical and Elec onic Enginee ing
314
ou he phan om hand ha seemed o b ing
empo a y elie om he phan om pain.
A e he membe s o he con ol g oup had
comple ed he s udy, i was shown ha all bu 2 o
he pa icipan s ound ha hey could ag ee wi h he
e ec c ea ed by he s anda d mi o box illusion.
They ag eed ha he sensa ion o isualising hei
e lec ion mo ing in unison wi h hei con olling
a m was qui e ealis ic. When using he Augmen ed
Reali y Mi o , 19 o he 22 subjec s ound ha he
compu e gene a ed image esponded and beha ed
in a con incing manne o ha o hei con olling
a m wea ing he glo e (2 o hose who we e unable
o expe ience a phan om sensa ion using he
Augmen ed Mi o we e he same subjec s ha
ound he s anda d mi o box illusion o ha e no
e ec ). 16 subjec s epo ed ha hey we e con en
ha he ca oon-like image beha ed easonably like
a ealis ic one when using he Augmen ed Reali y
mi o , bu many ound i o be a lo less con incing
han he image o a p ope a m.
Analysis o he accele ome e and EMG eadings
indica e ha 12 o he subjec s subconsciously
mo ed hei igh a m while pe o ming he asks.
We we e able o de ec an inc ease in he ac i i y o
he C3 si e in mos o he pa icipan s whene e
mo ion occu ed in he image. While we expec ed o
obse e his in he cases o people whose igh a m
mo ed behind he sc een/mi o , i was in e es ing o
no e an inc ease in ac i i y in some o hose whose
igh a m emained s a ic h oughou he
expe imen s.
7. DISCUSSION
To da e, we ha e un o una ely been unable o
ob ain a su icien numbe o igh -hand ampu ee
olun ee s o his s udy in o de o ca y ou a
p ope quan i a i e analysis. A p esen , ou
ampu ees ha e ied bo h ypes o mi o box. Th ee
o hese ound hem o ha e a bene icially
he apeu ic e ec on hei Phan om Limb and
assis ed hem in hei a emp o alle ia e Phan om
Limb Pain. The o he ampu ee ound any a emp o
inci e mo emen in his phan om ui less. P io o
expe imen a ion, none o he ampu ees had been
able o p oduce any la ge-scale olun a y mo emen
o hei phan om limbs. Howe e , by using bo h
mi o box echniques desc ibed he e o gi e
mo emen o hei phan om limb, h ee o he
ampu ees epo ed eeling a unique sensa ion ha
none o hem had expe ienced be o e. Whils hese
a e p elimina y esul s, hey a e none heless g ea ly
encou aging, in ha h ee o he ou ampu ees
es ed he Augmen ed Reali y mi o was able o
p o ide some o m o elie om phan om pain.
The ul ima e goal o his esea ch g oup is o c ea e
an en i onmen using AR echnology ha will allow
ampu ees o isually expe ience hei phan om limbs
i s hand. They should be able o see hei phan om
limbs as hey a e pe cei ed, as well as ha e he
abili y o con ol hei posi ion, mo ion and
o ien a ion. By ha nessing he po en ial o
Augmen ed Reali y echnology, compu e gene a ed
images o ampu ee’s expe ienced phan om limbs can
be c ea ed in such a way ha hey mo e and look
like hei own phan om expe ience. This b ings he
he apeu ic alue o he isual illusion c ea ed by he
mi o box o a wide ange o ampu ees. I may
hen be possible ha Augmen ed Reali y has he
po en ial o educe o maybe emo e phan om limb
pain by he isual eedback ecei ed om he
augmen ed en i onmen . This may help educe he
in luence o co ical emapping, e e se i , o maybe
e en p e en i om occu ing.
Acknowledgemen s
The au ho s hank D . Dei d e Desmond & P o .
Malcolm MacLachlan T ini y College Dublin and
D . Ga y Mc Da by & his G oup, o me ly o Media
Lab Eu ope o all hei help.
REFERENCES
[1] S. Wie -Mi chell: Lippinco ’s Magazine 8
(1871) 563.
[2] S. Wie -Mi chell: Inju ies o ne es, and hei
consequences. Philadelphia: J.B. Lippinco 1872
[3] V. S. Ramachand an and W. Hi s ein: B ain 121
(1998) 1603
[4] S. Sunde land: Ne es and ne e inju ies. 2
nd
ed.
Edinbu gh: Chu chill Li ings one 1978
[5] K. E. Li ings on: Jou nal o Neu osu ge y, 2
(1945) 25.
[6]P. Gallaghe and M. MacLachlan: Jou nal o
Heal h Psychology 6 (2001) 85.
[7] R. A. She man, C. J. She man and
L.Pa ke : Pain 18 (1984) 83.
[8] R. A. She man and C. J. She man: Am J Phys
Med 62 (1983) 227.
[9] J. Ka z and R. Melzack: Pain 43 (1990) 319.
[10] E. J. B owde and J. P. Gallaghe : Ann Su g
128 (1948) 456.
[11] S. A. Weiss and S. Fishman: Jou nal Abno mal
Soc Psycol, 66 (1963) 489.
[12] T. S. Jensen, B. K ebs, J. Nielson and P.
Rasmussen: Pain, 17 (1983) 243.
[13] J. H. Kaas, R. J. Nelson, M. Su , C. S. Lin and
M. M. Me zenich: Science 204 (1979) 521.
[14]M. M. Me zenich, R. J. Nelson, M. P. S yke ,
M. S. Cynade , A. Schoppmann A and J. M.
Zook:J Comp Neu ol 224 (1984) 591.
[15] W. Pen ield and T. L. Rasmussen:

The po ency o op ical and augmen ed eali y mi o boxes…
315
The ce eb al co ex o man: a clinical s udy o
localisa ion o unc ion. New Yo k: Macmillan;
1950
[16] H. Flo , T. Elbe , S. Knech , C. Wienb uch, C.
Pan e , N. Bi baume , W. La big and E. Taub:
Na u e 357 (1995) 482.
[17] N. Bi baume , W. Lu zenbe ge , P. Mon oya,
W. La big, K. Une l, S. Top ne , W. G odd, E.
Taub and H. Flo : Jou nal o Neu oscience 17
(1997) 5503.
[18] V. S. Ramachand an, M. S ewa and D.
Roge s-Ramachand an: Neu o epo 3 (1992)
583.
[19] P. M. Halligan, L. C. Ma shall, D. T. Wade, J.
Da ey and D. Mo ison: Neu o epo 4 (1993)
233.
[20] V. S. Ramachand an: P oc Na l Acad Sci USA
90 (1993) 10413.
[21] T. T. Yang, C. Gallen, B. J. Schwa z, F.
E. Bloom, V. S. Ramachand an and S. Cobb:
Na u e 368 (1994) 592.
[22] T. T. Yang, C. Gallen, V. S. Ramachand an, S.
Cobb, B. J. Schwa z and F. E. Bloom:
Neu o epo 5 (1994) 701.
[23] V. S. Ramachand an and D. Roge s-
Ramachand an: P oc R Soc Lond B Biol Sci 263
(1996) 377.
[24] A. W igh : Wellcome T us Sci A P ojec
(1997)
[25] H. Jaspe : Jou nal o Clinical Neu ophysiology,
7(2) (1980) 191.