Brain (1997),120,1217–1228

Mirror movements in X-linked Kallmann’ssyndromeII. A PET study

M. Krams,1 R. Quinton,2 M. J. Mayston,4 L. M. Harrison,4 R. J. Dolan,1,3 P.-M. G. Bouloux,2

J. A. Stephens,4 R. S. J. Frackowiak1 and R. E. Passingham1,5

1Wellcome Department of Cognitive Neurology, Institute of Correspondence to: Dr M. Krams, Wellcome Department ofNeurology,2Division of Endocrinology,3Academic Cognitive Neurology, Institute of Neurology, Queen Square,Department of Psychiatry, Royal Free Hospital School of London WC1N 3BG, UKMedicine,4Department of Physiology, University College,London,5Department of Experimental Psychology,University of Oxford, Oxford, UK

SummaryTo investigate the mechanism of mirror movements seen in significantly stronger. In the controls, significant increases in

rCBF were seen in the contralateral M1 during voluntaryX-linked Kallmann’s syndrome, we measured changes ofregional cerebral blood flow with H215O-PET. We studied six movement of either hand; a small ipsilateral M1 activation

was found in two out of six normal subjects when they movedright-handed Kallmann male subjects and six matched, right-handed control subjects during an externally paced finger their left hand. In a second experiment it was shown that, in

two out of two Kallmann subjects, passive movements of theopposition task. The analyses were done both on a singlesubject and a group basis. The Kallmann group showed a right hand resulted in left M1 activation that was similar to

the activation in the left M1 when subjects made mirrorstrong primary motor cortex (M1) activation contralateralto the voluntarily moved hand, but there was also a significant movements with their right hand. This suggests, but does not

prove, that the small but significant activation of thedegree of M1 activation ipsilateral to the voluntarily movedhand, i.e. contralateral to the mirroring hand. However, when ipsilateral M1 in Kallmann’s subjects may be due to sensory

feedback from the involuntarily mirroring hand.comparing contralateral to ipsilateral M1 activation, the M1activation contralateral to the voluntarily moved hand was

Keywords: mirror movements; Kallmann’s syndrome; PET; motor cortex; cerebral blood flow

Abbreviations: 1DI 5 first dorsal interosseus muscle; Ll5 left hemisphere response to voluntary left hand movement;Lr 5 left hemisphere response to voluntary right hand movement; M15 primary motor cortex; rCBF5 regional cerebralblood flow; Rl 5 right hemisphere response to voluntary left hand movement; Rr5 right hemisphere response to voluntaryright hand movement; S15 primary somatosensory cortex; SMA5 supplementary motor area; SPM5 statistical parametricmapping; XKS5 X-linked Kallmann syndrome

IntroductionFranz Kallmann first described the familial association of involuntary, simultaneous and similar movements of the

homologous contralateral finger. The same phenomenon mayhypogonadism with anosmia in three pedigrees (Kallmannet al., 1944). In one of these pedigrees, where a typically also be observed more proximally, sometimes even with

flexion/extension at the elbow joint. The mechanism ofX-linked mode of inheritance was evident, affected malesalso exhibited mirror movements. In our experience, mirror mirror movements is uncertain, with some authors suggesting

activity in an anomalous corticospinal tract projection (Conradmovements occur in 85% of males with X-linked Kallmannsyndrome (XKS) (Quintonet al., 1996a, b). Voluntary et al., 1978; Brittonet al., 1991; van der Lindenet al., 1991)

and others suggesting a reduction of transcallosal inhibitorymovements of any finger of one hand are associated with

© Oxford University Press 1997

1218 M. Kramset al.

activity resulting in simultaneous activation of both motor controls (age range 23–49 years) were studied for comparison(N1–N8). All subjects were right-handed as tested by thecortices (Forgetet al., 1986; Daneket al., 1992).

In the previous paper Maystonet al. (1997) present Edinburgh Handedness Questionnaire (Oldfield, 1971).The study involved the administration of 4.8 mSv effectiveneurophysiological data on 13 XKS subjects who display

mirror movements. Three findings are presented which are dose equivalent of radioactivity per subject, and was approvedby the Administration of Radioactive Substances Advisorytaken to suggest the presence of an abnormally developed

ipsilateral corticospinal projection. First, there was a short Committee of the Department of Health of the UK. Thesubjects gave informed written consent, and the study wasduration peak centred around time zero in cross-correlograms

constructed from multiunit EMG recordings from approved by the joint research ethics committee of theRoyal Postgraduate Medical School, Hammersmith Hospital,cocontracting left and right first dorsal interosseimuscles

(1DI). Secondly, unilateral focal magnetic brain stimulation London and by the ethics committee of the Royal FreeHospital and School of Medicine, London.of primary motor cortex (M1) induced EMG responses in

homologous left and right hand muscles of similar shortlatency. Thirdly, stimulation of the digital nerves of the indexfinger of one hand caused modulation of ongoing EMG inInvestigationsthe opposite hand in most subjects. All subjects underwent a neurological examination to exclude

The first experiment of the present study used PET toany relevant neurological symptoms other than mirroring inmeasure regional cerebral blood flow (rCBF) in six of thesethe XKS group.13 XKS subjects with mirror movements. If the mirror All subjects had cranial MRI to obtain T1-weighted scans.movements are directed by M1 ipsilateral to the mirroringA 3D reconstruction and editing of the MRI data werehand, one might expect to find no activation of M1 performed using ANALYZE (Robb and Hanson, 1991).contralateral to that hand. The subjects were scanned while All but two subjects were scanned using a CTI modelperforming simple voluntary movements with either the left953B-PET scanner (CTI Inc, Knoxville, Tenn., USA) withor the right hand. We deliberately studied simple distalcollimating septa retracted. Subjects K4a and K12 weremovements because we wished to minimize the chances ofscanned on a Siemens EXACT HR1. For each scan, subjectsseeing activation ipsilateral to the voluntarily moved handreceived a 20-s intravenous bolus of H2

15O through a cubitalmerely due to the complexity of the movement executed. Infossa vein of the left arm. Twelve consecutive PET scansnormal subjects, ipsilateral activation of M1 has been reportedwere collected at 10-min intervals, each over a period of 2either when proximal muscles are involved (Stephanet al., min, beginning with a 30-s background scan before delivery1995) or when the movements are complex (Raoet al., 1993). of the bolus. The integrated radioactivity counts accumulated

However, there are two problems. First, sensory feedbackover the 90-s acquisition period, corrected for background,from the mirroring hand should be apparent in primarywere used as an index of rCBF. With a field of view of 10.8somatosensory cortex (S1) contralateral to that hand. Thus itcm in the z-plane, the subjects were positioned so as tois essential to distinguish between activation of M1 and S1.include the top of the brain, including all of the supplementaryA high-resolution PET camera was therefore used in 3-Dmotor area (SMA), and much of the cerebellum.mode, and the data for individual subjects were coregistered Surface EMG was recorded simultaneously from the leftonto the individual’s MRI scan for each subject. and right 1DI in all subjects with the exception of the normal

The second problem is that it has been shown in monkeyssubjects N1, N2 and N3. During the rest condition and duringthat the activity of cells in M1 can be modulated by passive movements high gain (50µV/cm) was used. For thecontralateral passive hand movements (Cheney and Fetz,XKS subjects, the EMG was rectified and averaged, and1984). In a second experiment we therefore measured thetime-locked to the beginning of the voluntary movement; theeffect of passive movements. We compared activation of theratio of involuntary to voluntary activity was calculated usingleft M1 in the XKS subjects when they moved their left handthe areas of the rectified EMGs for voluntarily intendedvoluntarily, mirroring with their right, with activation of the movements in both the right and left hands.left M1 when the right hand was passively moved so as tomimic the mirror movements.

Experiment 1: unilateral voluntary fingermovements

Methods ParadigmSix XKS subjects exhibiting mirror movements and sixSubjects

We studied eight XKS males (age range 16–48 years) normal controls were investigated. Prior to scanning, allsubjects were trained to perform a simple finger oppositionexhibiting mirror movements. All subjects had already been

studied electrophysiologically; see Maystonet al. (1997). task. The movements were paced by an electronic metronomeat 1 Hz. The task involved brisk phasic opposition of indexThe numbering of the subjects (K2, K4a, K5, K6, K8, K9,

K10, K12) is identical to that used in that study. Eight normal finger and thumb, with the tips of index finger and thumb

Mirror movements in XKS: II 1219

only briefly touching before extending again. There were of XKS subjects and the activation (movement versus rest)for the controls. This comparison was carried out for voluntarythree conditions: (A) voluntary finger–thumb opposition of

the right hand; (B) voluntary finger–thumb opposition of the movements with each hand.left hand; (C) (baseline) subjects were asked to relax and notto perform, or think about, any finger movements (the pacingActivation contralateral to the voluntarily movingtone was again presented at 1 Hz). and mirroring hand.For each M1 area, the size of

activation contralateral to the voluntarily moved hand wascompared directly to activation contralateral to the mirroring

Data analysis hand. For example, a comparison was made between theAnalysis of data was performed on SUN SPARC 20activation of the left M1 when subjects moved their rightworkstations (Sun Microsystems Inc., London, UK) usinghand voluntarily and when they moved their left handSPM-95 (Wellcome Department of Cognitive Neurology, voluntarily (and their right involuntarily). This techniqueLondon, UK) (Fristonet al., 1995a, b). After normalization, allows an assessment of which hemisphere is more stronglythe PET data-set extended from –32 mm below the AC–PCactivated when there is bilateral activation of homologous(anterior–posterior commissure) line to172 mm above it. areas.Calculations and image-matrix manipulations were performedin PRO MATLAB (Mathworks Inc., New York, USA). Single subject analysis.For each individual, a com-

For each subject, all 12 rCBF scans were realigned toparison was made of the rCBF values during movement andcorrect for head movement during scanning. A mean imageat rest. This was done for voluntary movement of eachof these 12 scans was then used to coregister PET data ontohand. For each M1 area, the activation contralateral to thethe same individual’s MRI scan. PET and MRI data werevoluntarily moved hand was then compared directly to thenormalized into Talairach space (Talairach and Tournoux,activation contralateral to mirror movements. This allowed a1988). A smoothing filter of 12 mm was used to accommodatecomparison of the size of activation of M1 contralateral tointer-subject differences in gyral anatomy, and to optimizevoluntary movements and with that contralateral to mirrorthe signal-to-noise ratio. Differences in global activity within movements.and between subjects were removed by analysis of covariance.Using the t statistic on a voxel-by-voxel basis, statisticalparametric mapping (SPM{t}) maps were generated with Experiment 2: M1 activation during passiveareas of activation ofP , 0.001 for the group andP , 0.01

finger movementsfor single subjects. Oura priori hypothesis was that, ifTwo right-handed XKS subjects (K4a and K12) and twovoluntary and mirror movements are controlled by the sameright-handed normal subjects (N7 and N8) were studied.M1, there should be only unilateral M1 activation during

unilaterally intended finger movements. In view of thisexplicit prior hypothesis no correction for multiplecomparisons was made. Paradigms

In this experiment we compared active versus passive fingerAfter coregistration of the single-subject high resolution3D PET data onto individual MRI scans, the central sulcus movements. Prior to scanning, subjects were trained to relax

while the experimenter passively moved their right indexwas determined on a transverse cut atz 5 160 mm, as thesulcus which lay between the marginal segment of the finger and thumb, mimicking the brisk active movement of

finger–thumb opposition. Subjects were further instructed notcingulate sulcus (posteriorly) and the paracentral sulcus(anteriorly). To determine, in single subjects, whether a peak to think about the movements. Surface electrodes on 1DI

monitored the EMG activity. The presence of any EMGwas localized either in M1 or S1 the central sulcus wasfollowed on transverse planes down to the level of the peak activity could then be recognized by the subjects, as EMG

biofeedback was provided in the training phase via a loudof the activation. M1 was defined as the anterior bank of thecentral sulcus, S1 as the posterior bank of the central sulcus. speaker.

During PET scanning the conditions for the XKS subjectswere: (A) voluntary active finger–thumb opposition of theleft hand; (B) passive finger–thumb opposition of the rightPlanned comparisonshand, by the investigator, so as to mimic the involuntarymirroring in the right occurring in condition A; (C) (baseline)Activation for each group.For both the XKS and normal

groups a comparison was made of the rCBF values during the investigator held the index finger and thumb of thesubject’s right hand, but without moving them.movement and at rest. This was done for movement of

each hand. During PET scanning the conditions for the normal subjectswere: (A) voluntary active finger–thumb opposition of theright hand; (B) passive finger–thumb opposition of the rightInter-group comparison.A direct comparison was made

between the activation (movement versus rest) for the group hand so as to mimic the movement in A; (C) (baseline) the

1220 M. Kramset al.

investigator held the index finger and thumb of the subject’s movement of the left hand are to be found in column Rl(activation contralateral to the voluntarily moved hand) andright hand, but without moving them.

During conditions B an investigator moved the subject’s Ll (activation ipsilateral to the voluntarily moved hand).Independent peaks in M1 and S1 are listed as such in theright index finger and thumb, mimicking the finger–thumb

opposition movements which occurred in condition A. In the tables. Whenever there is only a single peak in either M1 orS1, but the activation extends into the adjacent area, thenormal subjects, the finger tips of thumb and index finger

touched during both conditions (A and B). In the XKS coordinates of the peak are shown in the appropriate area inthe table and also repeated (in square brackets) for the areasubjects, however, the finger tips were not brought to contact

during the passive condition, since subjects never brought into which the activation extends.Both groups showed strong contralateral M1 and S1the tips of thumb and index finger into contact when mirroring.

During condition B the investigator held the subject’s fingers activation. An ipsilateral activation (contralateral to themirroring hand) was found only in the XKS group duringlaterally at the distal interphalangeal joint. The passive

movements were paced by the metronome at 1 Hz. One voluntary movements of either hand.Both groups showed strong activation of the SMA,normal subject was only studied in conditions B and C.

extending ventrally into anterior cingulate regions. The SMAactivation was mostly contralateral to the voluntarily movedhand. Additionally there was some ipsilateral activation inData analysisthe XKS group during voluntary movements of the left hand.The image and data analysis was as for Experiment 1 with

The lateral premotor area (Brodmann area 6) was bilaterallythe following changes. This experiment was performed lateractivated in the normal group during movements of eitherthan Experiment 1, and therefore SPM 96 (Wellcomehand. The XKS group showed bilateral activation duringDepartment of Cognitive Neurology, London, UK; Fristonvoluntary movements of the left hand. During right handet al., 1995b) had become available and could be used. SPMmovements, significant foci of activation were only present96 has the advantage that the fit of the PET to the MRI isin the right lateral premotor area. Activation of the lateralmore reliable, and the data are examined in the stereotaxicpremotor cortex extended ventrally into frontal opercular andspace as defined by Evanset al. (1991, 1993). Afterinsular regions.normalization, the PET data-set for the two XKS subjects

There was strong cerebellar activation ipsilateral to theextended from –52 mm below the AC–PC line to184 mmvoluntarily moved hand in both groups, and this was true forabove it and from –30 mm to178 mm for the twovoluntary movements of either hand. Additional contralateralnormal subjects. SPM{t} maps were generated with areas ofactivation occurred during voluntary movements of the rightactivation whereP , 0.005 for single subjects.hand in the normal group and the left hand in the XKS group.

In the normal group there was a focus in the left putamenduring voluntary movements of either hand, as well as aPlanned comparisonsfocus in the right putamen during movements of the leftThe following planned comparisons of rCBF values werehand. For the XKS group there was a single focus in thecarried out for single subjects: voluntary finger movementright putamen during voluntary movement of either hand.versus baseline (A versus C); passive finger movement versus

baseline (B versus C); voluntary finger movements versusInter-group comparison.For the following areas thepassive finger movements (A versus B).XKS group showed stronger activation than the normalgroup: M1/S1 ipsilateral to the voluntarily moved hand, i.e.contralateral to the mirroring hand (Ll coordinates5 –30,Results–28, 148, Z 5 3.72; Rr coordinates5 132, –28, 152,

Experiment 1: unilateral voluntary finger Z 5 4.99); right cerebellar vermis for voluntary movementmovements of either hand (Rl coordinates5 108, –60, –12,Z 5 3.76;

Rr coordinates5 106, –50, –08, Z5 3.62); right putamenPETActivation for each group.Table 1 reports the results for voluntary movements of the right hand (Rr coordinates5

124 –16104, Z 5 3.58).of the group analysis for each group when comparingmovement versus rest. The columns are organized accordingto hemisphere (upper case letters: L5left hemisphere, R5 Activation contralateral to the voluntary and

mirroring hand.This comparison relates the activation inright hemisphere) and the hand that was moved voluntarily(lower case letters: r5 moving right hand, l5 moving left each hemisphere of XKS subjects when the activation of that

hemisphere was contralateral to voluntary movement and thehand). Thus foci of significant activation during voluntarymovement of the right hand are to be found in column Lr activation of the same hemisphere when the activation was

contralateral to mirror movements. Differences were found(activation contralateral to the voluntarily moved hand) andRr (activation ipsilateral to the voluntarily moved hand). only for M1/S1 (P , 0.001). For the left hemisphere, M1/

S1 were more activated when contralateral to the voluntarilyLikewise, foci of significant activation during voluntary

Mirror movements in XKS: II 1221

Table 1 Areas of activation (P , 0.001) andZ-scores in normal and XKS subjects, movement compared with rest

Area Lr Rl Ll Rr

Z-score x, y, z Z-score x, y, z Z-score x, y, z Z-score x, y, z

Normal subjectsM1 7.50 –46, –26,144 7.70 132, –28,148 n.s. n.s.S1 8.54 –30, –32,152 7.70 [132, –28,148] n.s. n.s.SMA/anterior cingulate 5.17 –06, –08,148 5.93 102, –10,148 n.s. n.s.Lateral premotor 3.43 –56,104, 120 3.98 150, –08,136 3.52 –54, –04,132 3.69 148, –08,136Insula/frontal operculum 3.43 [–56,104, 120] 4.50 134, –02 00 3.48 –46, –02,108 n.s.Cerebellum 3.86 –16, –66, –20 n.s. 6.55 –10, –62, –16 6.07 112, –60, –16Basal ganglia 4.64 –26, –04,108 3.77 128, –10,104 4.66 –26, –04,108 n.s.

XKS subjectsM1 6.29 –36, –26,148 7.43 [134, –30,152] 4.00 –36, –22,148 4.84 132, –26,152S1 6.29 [–36, –26,148] 7.43 134, –30,152 4.00 [–36, –22,148] 4.84 [132, –26,152]SMA/anterior cingulate 4.94 –04, –12,148 3.73 108, –08,148 4.10 –06, –14,148 n.s.Lateral premotor 3.35 –56,102, 128 3.96 154, 104, 116 3.56 –56 00,132 n.s.Insula/frontal operculum n.s. 3.69 140, 104, 112 n.s. n.s.Cerebellum n.s. 5.8 102, –64, –12 5.04 –08, –56, –08 5.85 104, –62, –12Basal ganglia n.s. 5.76 124, –16,104 n.s. 5.351 24, –16,104

The coordinates andZ-scores refer to the most significant focus in that area. Where activation extends into two areas, as for M1 and S1, but there is asingle focus, the coordinates of the peak are given, and repeated in square brackets for the area into which the activation extends. The Talairachcoordinates (Talairach and Tournoux, 1988) are given in mm for the maximally significant pixel in each area:x 5 lateral displacement from the midline,negativity leftwards;y 5 anteroposterior displacement relative to the anterior commissure, posterior negative;z 5 vertical position relative to the AC–PCline, negativity downwards; n.s.5 not significant. Columns are organized according to hemisphere (L5 left hemisphere; R5 right hemisphere) and thehand that was voluntarily moved (l5 left hand; r5 right hand). Thus activations contralateral to the voluntarily moved hand are in the Lr and Rlcolumns, and ipsilateral activations are in the Ll and Rr columns.

moving hand (coordinates5 –36, –32,148, Z 5 3.84). For In four out of six XKS subjects with mirror movements(K2, K6, K9 and K10) the predominant activation lay in thethe right hemisphere, M1/S1 were also more activated when

contralateral to the voluntarily moving hand (coordinates5 M1 contralateral to the voluntarily moved hand; this can beseen from the comparisons made in the two columns on the134, –34,152, Z 5 5.18).right of Table 1. In all the XKS subjects there was activationof the M1 ipsilateral to the voluntarily moved hand, that isIndividual subject activation.For the individual data,

the significance level isP , 0.01; this level was chosen to contralateral to the mirroring hand. However, the degree ofthis activation differed greatly between subjects. For Subjectmaximize the chance of finding ipsilateral M1 activation in

the individuals in the normal group. All normal subjects K5 there was a large right M1 activation whichever hand hemoved voluntarily and only a small left hemispheric activationshowed strong activation of M1 contralateral to the voluntarily

moved hand. Additionally there were small but significant higher up in M1. For Subject K8 both the left and right M1were strongly activated whichever hand was voluntarilyipsilateral M1 activation in Subjects N2 and N6 when they

moved their non-dominant left hand. moved.Figure 1 shows the activation for the individuals in theTable 2 lists the foci of activation in M1 for the individual

XKS subjects. The layout of Table 2 follows the conventions XKS group when voluntarily moving their right or left hand.Transverse sections have been cut through M1 at the leveloutlined for Table 1. However, there are two additional

columns: Lr–Ll and Rl–Rr. These columns list significant of the most significant focus in M1. The PET images for theindividuals are coregistered onto their MRI scans.foci when comparing the activation of each hemisphere when

it is contralateral to voluntary movements (Lr and Rl, All normal subjects showed S1 activation contralateral tothe voluntarily moved hand. None of the normal subjectsrespectively) with the activation of the same hemisphere

when it is contralateral to mirror movements (Ll and Rr, showed ipsilateral S1 activation.Table 3 shows the foci of S1 activation for the individualrespectively).

Each activation was checked for its anatomical location XKS subjects. There were distinct S1 and M1 foci in all buttwo XKS subjects (K2 and K9); in these two, M1 activationby coregistering the individual PET image with the individual

MRI. This allowed a distinction to be drawn between extended into S1. Five out of six XKS subjects showedbilateral S1 activation. In Subject K5 no S1 activation wasactivation in M1 and S1 by determining whether the focus

of activation was anterior or posterior of the central sulcus. found in the left hemisphere. In all other subjects there were

1222 M. Kramset al.

Tab

le2

Act

ivat

ion

ofM

1(P

,0.

01)

inX

KS

subj

ects

Subj

ect

Lr

Rl

Ll

Rr

Lr–

Ll

Rl–

Rr

Z-s

core

x,y,

zZ

-sco

rex,

y,z

Z-s

core

x,y,

zZ

-sco

rex,

y,z

Z-s

core

x,y,

zZ

-sco

rex,

y,z

(z-e

xten

t)(z

-ext

ent)

(z-e

xten

t)(z

-ext

ent)

(z-e

xten

t)(z

-ext

ent)

K2

4.89

–44,

–26,

140

5.13

140

,–2

2,1

443.

85–4

4,–2

6,1

404.

101

40,

–20,

148

3.10

–44,

–26,

140

4.39

144

,–2

2,1

44(1

36to

156

)(1

40to

152

)(1

36to

152

)(1

44to

152

)(1

40)

(140

to1

52)

K5

3.11

–14,

–18,

160

3.10

132

,–2

6,1

522.

92–1

2,–1

8,1

602.

721

32,

–26,

152

n.s.

n.s.

(160

)(1

40to

156

)(1

60)

(148

to1

52)

K6

3.62

–38,

–22,

144

3.25

146

,–1

8,1

402.

61–3

2,–2

4,1

482.

381

30,

–20,

152

3.14

–44,

–22,

140

3.4

134

,–1

4,1

48(1

40to

160

)(1

40to

160

)(1

48to

160

)(1

48to

156

)(1

40to

148

)(1

44to

148

)

K8

4.75

–38,

–18,

148

3.71

136

,–1

8,1

484.

53–3

8,–1

6,1

483.

501

30,

–20,

152

2.2*

–40,

–24,

148

n.s.

(136

to1

56)

(140

to1

56)

(140

to1

52)

(144

to1

56)

(148

to1

52)

K9

4.68

–52,

–22,

140

3.59

140

,–2

2,1

443.

85–5

2,–2

2,1

402.

51[1

36,

–32,

148

]4.

33–5

0,–2

6,1

483.

121

40,

–22,

144

(136

to1

52)

(140

to1

64)

(136

to1

40)

(148

)(1

40to

168

)(1

44to

164

)

K10

4.84

–42,

–22,

144

3.71

146

,–2

6,1

483.

9–4

6,–2

2,1

483.

36[1

30,

–30,

156

]4.

25–4

2,–2

2,1

444.

021

44,

–22,

156

(136

to1

60)

(140

to1

64)

(140

to1

52)

(152

to1

60)

(136

to1

60)

(136

to1

64)

Con

vent

ions

asin

Tabl

e1.

Col

umns

Lr–

Ll

and

Rl–

Rr

list

sign

ifica

ntfo

ciw

hen

com

pari

ngM

1ac

tivat

ion

cont

rala

tera

lto

volu

ntar

ym

ovem

ents

with

that

cont

rala

tera

lto

mir

ror

mov

emen

ts,

inth

ele

ftan

dri

ght

hem

isph

eres

,re

spec

tivel

y.* P

,0.

05.

The

exte

ntof

the

activ

atio

nin

the

z-ax

isis

give

nin

brac

kets

(z-e

xten

t).

Mirror movements in XKS: II 1223

Fig. 1 Pet activation in the individuals in the XKS group during voluntarymovements of their left or right hand. Transverse sections have been cut through M1at the height of the focus in M1 with the highestZ-score. Areas of activation arecoregistered onto the individual’s MRI scan.

strong contralateral S1 activation. Additionally all but one was no EMG activity during rest or in the resting hand inSubject N5. However, low amplitude EMG activity was(K5) XKS subjects showed S1 activation ipsilateral to the

voluntarily moved hand (contralateral to the mirroring hand). found in the resting hand in Subjects N4 and N6 when theymoved their left hand voluntarily. There was also ongoingThe degree of ipsilateral activation varied between subjects.low amplitude EMG during rest in both hands in Subject N6.

XKS subjects.Table 4 lists the ratio of the EMGs in 1DIEMG recordingof the mirroring versus voluntarily moved hand. It also listsin how many runs there was low amplitude EMG duringNormal controls.Strong and high amplitude EMG in the

voluntarily moved hand was found during all runs. There rest. It can be seen that all the XKS subjects showed mirror

1224 M. Kramset al.

Table 3 Activation of S1 (P , 0.01) in XKS subjects

Subject Lr Rl Ll Rr

Z-score x, y, z Z-score x, y, z Z-score x, y, z Z-score x, y, z(z-extent) (z-extent) (z-extent) (z-extent)

K2 4.89 [–44, –26,140] 2.92 148, –38,140 3.85 [–44, –26,140] 4.10 [140, –20,148](140 to 152) (136 to 152) (136 to 144) (144 to 148)

K5 n.s. 2.60 126, –30,140 n.s. 2.55 126, –30,140(140 to 144) (140)

K6 5.04 –24, –36,156 4.54 134, –36,152 3.97 –26, –32,156 2.87 132, –34,152(148 to 160) (148 to 156) (152 to 156) (152)

K8 3.78 –48, –26,140 4.38 150, –20,124 3.12 –48, –26,140 3.62 150, –20,124(132 to 152) (120 to 132) (136 to 148) (120 to 136)

K9 4.47 –42, –32,156 4.42 134, –38,160 3.85 [–52, –22,140] 2.51 136, –32,148(136 to 160) (140 to 164) (140 to 144) (148)

K10 3.48 –40, –32,136 5.38 130, –30,156 3.7 –38, –36,128 3.36 130, –30,156(132 to 144) (144 to 160) (124 to 136) (152 to 160)

Conventions as in Table 1. The extent of the activation in thez-axis is given in brackets (z-extent).

Table 4 Ratio of involuntary versus voluntary EMG in In Subjects K4a and K12, the M1 activity in the leftXKS-subjects during PET scanning hemisphere was similar when the right hand was involuntarily

mirroring and when it was passively moved and a directSubject Right hand active Left hand active EMG during restcomparison of active versus passive did not reveal anyI/V (n 5 4) I/V (n 5 4) (n 5 4)differences. In Subject K4a the extent of left hemispheric

K2 0.07 0.09 0/4 M1 activation was small during active movement of theK5 1.18 0.60 2/4 bilateral left hand.K6 0.34 0.53 3/4 mostly left

Figure 2 shows the PET activity for Subjects N8, K4a andK8 0.38 0.21 4/4 bilateralK12 during active and passive finger movements coregisteredK9 0.55 0.45 3/4 bilateral

K10 0.02 0.03 1/4 mostly right onto the individual’s MRI scans.None of the subjects showed any EMG activity in either

The ratios of involuntarily (I, mirroring) versus the voluntarily hand during rest or with passive hand movements, with the(V) EMG in homologous muscles (1DI) are given in columns 2exception of one run of passive hand movements in Subjectand 3. Column 4 lists the number of rest periods (out of four)K4a, which was excluded from the analysis due to ongoingduring which low amplitude EMG was registered in either hand.activity in the left 1DI.

movements, though there was considerable variation in thedegree of involuntary compared with voluntary activity. InSubject K5 the EMG activity was stronger in the mirroring Discussionhand when the subject moved his right hand voluntarily.There are five main findings. First, in the single subjectsubjects K2 and K10 showed only minor degrees of EMGanalysis significant activation of M1 in all the XKS subjectsactivity in the mirroring hand. was found ipsilateral to the voluntarily moved hand, i.e.

contralateral to the mirroring hand. M1 activation ipsilateralto the voluntarily moved hand occurred in only two out of

Experiment 2: M1 activation during passive six normal subjects. Ipsilateral activation of M1 has beenreported in other studies of normal subjects using PET andfinger movements

Table 5 lists significant activation in S1/M1. All subjects fMRI (functional MRI) (Raoet al., 1993; Stephanet al.,1995), but Raoet al. (1993) report finding ipsilateralshowed M1 activation contralateral to the passively moved

right hand. In Subject N8, the degree and extent of M1 activation only for complex movements and Stephanet al.(1995) in movements involving the proximal musculature.activation was similar in active and passive finger movements;

a direct comparison of active versus passive did not reveal Secondly, when the XKS group was compared with thenormal group, there was a significantly greater activation inany significant difference in M1 activation in this subject.

Both XKS subjects tested showed bilateral S1/M1 S1/M1 ipsilateral to the voluntarily moved hand in the XKSgroup. The groups also differed in that there was moreactivation during active finger movements of the left hand.

Mirror movements in XKS: II 1225

Table 5 Activation in M1 and S1 (P , 0.005) during active and passive finger movements in individual subjects

Finger movements Left M1 Left S1 Right M1 Right S1

Z-score x, y, z Z-score z, x, y Z-score x, y, z Z-score x, y, z(z-extent) (z-extent) (z-extent) (z-extent)

N7Passive right 4.44 –44, –14,150 4.75 –62, –22,148 n.s. n.s.

(138 to 170) (138 to 172)

N8Passive right 4.91 –46, –22,152 4.91 [–46, –22,152] n.s. n.s.

(134 to 164) (134 to 164)

Active right 5.19 –46, –22,152 5.19 [–46, –22,152] 3.88 134, –28,142 n.s.(134 to 168) (134 to 166) (134 to 142)

Active versus passive n.s. n.s. n.s. n.s.

K4aPassive right 3.67 [–52, –28,146] 3.67 –52, –28,146 n.s. n.s.

(142 to 148) (144 to 150)

Active left 3.02 –46, –22,148 3.81 –52, –32,150 3.33 128, –18,160 3.33 [128, –18,160](142 to 148) (146 to 156) (152 to 164) (152 to 162)

Active versus passive n.s. 3.29 –52, –34,156 n.s. 3.45 154, –32,142(152 to 158) (134 to 144)

K12Passive right 3.59 –32, –22,168 4.20 –40, –32,164 n.s. n.s.

(166 to 170) (158 to 170)

Active left 3.80 –32, –20,168 3.40 –48, –26,166 3.48 [124, –30,152] 3.48 124, –30,152(166 to 174) (164 to 168) (150 to 160) (152 to 160)

Active versus passive n.s. n.s. 3.79 128, –22,156 3.38 124, –30,152(152 to 176) (152 to 162)

Coordinates are in stereotaxic space as defined by Evanset al. (1991, 1993),Z-score and extent in thez-plane refer to the focus with the mostsignificant activation in that area. Where the activation extends into two areas, e.g. M1 and S1, but there is a single focus, the coordinates of thepeak are given for wherever the peak is localized and repeated in square brackets in the area into which the activation extends. The extent of theactivation in thez-axis is given in brackets (z-extent).

activation for the XKS group in the right putamen and the bank of cingulate cortex have been shown to induce mirrormovements in monkeys when they perform bimanual tasksright cerebellar vermis. It is not clear how these differences

could explain the mirror movements. We assume that they (Brinkman 1984). However, if mirroring in XKS subjectswas caused by a disturbance of the SMA and cingulateresult from the differences in ipsilateral S1/M1 activation.

Thirdly, for the XKS group, activation in S1/M1 was cortex, one would have predicted a robust difference betweenthe groups in these areas.greater contralateral to the voluntarily moved hand than

contralateral to the mirroring hand, whichever hand was Fifthly, passive movement of the right hand was shown toactivate M1 contralateral to the passively moved hand in allused. This was also the case with four out of six XKS

individuals (K2, K6, K9 and K10) in the single subject XKS and normal subjects studied in the later experiment.This has important consequences for the interpretation of ouranalysis. For a fifth individual (K8) this only held for the

left hemisphere and only if a lower significance level was data and will be discussed first.accepted (P , 0.05).

Fourthly, at a significance level ofP , 0.001 there wasno difference between the XKS and normal group in theActive versus passive movements

In Experiment 2 we have shown that M1 is activated duringSMA and the underlying cingulate cortex. This is importantbecause there are dense callosal connections between the left passive movement of the contralateral hand. This has also

been shown in previous studies (Zeffiro and Hallett 1992;and right SMA and the left and right cingulate cortex (Rouilleret al. 1994), and unilateral lesions of the SMA and upper Bernardet al., 1996). The second of these studies used fMRI

1226 M. Kramset al.

Fig. 2 Pet activation in Subjects N8 (normal), K4a and K12, during active andpassive finger movements. Transverse sections have been cut through M1. Areas ofactivation are coregistered onto the individual’s MRI scan.

and clearly shows that the activation includes M1. In our subjects. Their table 4 has been arranged according to the sizeof the ipsilateral response projection as revealed using focalstudy, direct comparisons of M1/S1 activation in the left

hemisphere showed that for Subject N8 there was no magnetic brain stimulation.The single subject analysis of PET data has shownsignificant difference in M1 activation between active and

passive right finger movements. In both XKS subjects studied, considerable inter-subject differences within XKS patients, inparticular with regard to degrees of contra- and ipsilateral M1the M1/S1 activation contralateral to the mirroring hand was

similar to that seen when the same hand was passively moved. activation. The order seen in the electrophysiological data,however, does not correspond to any order that can be seen inIt seems possible that in subjects with mirror movements the

M1 activation ipsilateral to the voluntarily moved hand (i.e. the PET data, nor is the amount of mirroring, as recorded byEMG during PET scanning, reflected in the PET activationcontralateral to the mirroring hand) can, at least in part, be

accounted for by sensory feedback from the mirroring hand. data.For example, it can be inferred from the electrophysiologicalIt is important to note that the M1/S1 activation

contralateral to the passively moved hand was more extended findings that Subject K2 has a more pronounced large-diametercorticospinal projection ipsilaterally than contralaterally. Thisin the normal subjects. This may be explained by the

additional cutaneous afferent input due to the touching of leads one to predict that the PET study would reveal a strongerM1 activation ipsilateral to the voluntarily moved hand.the fingertips, which was present in the active and passive

movements of the normal subjects, but not in the mirroring However, this was not the case. Subject K2 clearly showsbilateral M1 activation, independent of which hand isor passive mimicking of mirroring in the XKS subjects.voluntarily moved. Furthermore, the contralateral M1activation seems to be more pronounced, independent of whichhand is being used. One explanation is that this subject doesDifferences between XKS subjects

Mayston et al. (1997) report significant inter-subject use both motor cortices to initiate finger movements.Alternatively the M1 activation may be a combination ofdifferences in their electrophysiological findings on XKS

Mirror movements in XKS: II 1227

afferent and efferent activity. As this subject showed only weak Thirdly, in the study by Maystonet al. (1997) the data arederived from several sources. Focal magnetic brain stimulationmirroring, one would expect a stronger sensory feedback from

the voluntarily moving hand. This may account for the more was used to evoke excitation of motor neurons artificially, andcutaneo-muscular reflexes were recorded following excitationpronounced contralateral M1 activation.

In Subject K5, the extent of the right M1 activation is greater of digital nerves. In this study, on the other hand, the subjectswere intending voluntary movements of one hand. Maystonthan that of the left M1, regardless of which hand is being

voluntarily moved. The M1 activation is also strikingly similar et al. (1997) also correlate the EMG activity from both handswhile subjects were engaging in voluntary movement.duringvoluntary movementsof eitherhand. Anothersurprising

finding is the lack of left S1 activation in this patient, who had However, while the results indicate that the responses of thetwo hands can be correlated in time, they do not define thestrong mirror movements. This is also the only subject whose

involuntary mirror movements of the left hand are more spatial location of the common generator, although M1 remainsthe most likely site.pronounced than the voluntary movements of the right hand.

Data from focal magnetic brain stimulation for Subject K8suggest that the ipsi- and contralateral corticospinal largediameter projections are similar. In this subject who had fairlyControl of the mirroring hand

Cohenet al. (1991) have previously reported a PET study instrong mirror movements, the activation of M1 is similar forthe left and right M1, regardless of whether the activation is which they found bilateral activationof thesensorimotor cortex

in two patients with congenital mirror movements. Our PETcontralateral to the voluntarily moved hand or the mirroringhand. The M1 activation may again reflect a combination of study has revealed that XKS subjects with mirror movements

have indeed different degrees of bilateral activation of M1sensory feedback and motor control.According to the electrophysiological findings, Subjects K9 when performing a unilaterally intended movement. One

interpretation of this result could be that mirror movements inand K10 have a predominantly contralateral, large-diametercorticospinal projection. The more pronounced contralateral these subjects result from a simultaneous activation of the left

and right motor cortex, as suggested by Shibasaki and NagaeM1 activation in these subjects could be seen to be consistentwith the electrophysiological data. Alternatively, the stronger (1984). They examined movement-related cortical potentials

in an XKS subject and found a premovement negative responsecontralateral M1 activation could be explained by additionalcutaneous afferent feedback resulting from the touching of the bilaterally in response to intended unilateral hand movement.

They then argued that mirror movements were generated byfingertips of the voluntarily moved hand. The fingertips of themirroring hand did not touch. However, whereas Subject K9 unintended excitation of the primary motor cortex opposite to

the involuntarily moved hand. Daneket al. (1992) suggestedexhibits strong mirroring, Subject K10 only mirrors slightly.This difference is not expressed in the PET findings. that there may be a lack of transcallosal inhibition in XKS

subjects with mirror movements. Such a hypothesis wouldpredict bilateral activation of M1 as found in the present study.More recentlyMayeretal. (1995) comparedmovement-relatedComparison of PET and physiology

Three problems arise when comparing PET activation data and potentials in patients with autosomal dominant (i.e. non-Kallmann syndrome) mirror movements and normal subjectsthe results of neurophysiological experiments.

First, the present study demonstrates that M1 activation while they were executing unilaterally intended fingermovements. Whereas there was no difference in premovement-occurs when the contralateral hand is moved passively. It has

been shown directly that there are cells in the primate M1 related cortical potentials, movement-related potentials aroundthe onset of the EMG recorded from both hands was bilateralthat are responsive during passive movement (Lemon, 1981;

Cheney and Fetz, 1984; Andersson, 1995). Thus, in the XKS in patients with mirror movements, but only contralateral to thevoluntarilymovedhand in thenormalsubjects.Theyconcludedgroup, M1 activation ipsilateral to the voluntarily moved hand

is also explicable in terms of sensory feedback from the that the ipsilateral cortical activation around movement onsetmay be associated with a cortical mechanism trying toinvoluntarily mirroring hand.

Secondly, it is believed that the changes in rCBF measured compensate for abnormal ipsilateral corticospinal pathways insubjects with persistent mirror movements.in PET relate to changes in the activity in cell terminations

(Jueptner and Weiller, 1995). If this is so, activation seen in Maystonet al.(1997) have presented evidence that suggeststhat, inat least someXKSsubjects, themirrormovementscouldM1 probably represents the summated afferent activity from

regions that project to M1; these include S1, parietal area 5, result fromactivity inanabnormallydeveloped ipsilateral tract.The PET data could be interpreted as being consistent with thispremotor cortex, the SMA and ventral thalamus (Muakkassa

and Strick, 1979; Ghoshet al., 1987; Matelli and Luppino, hypothesis if it is supposed that the M1 activation contralateralto the mirroring hand is the result of the mirror movements1993). Thus, whereas magnetic stimulation excites the

pyramidal output cells either indirectly or directly but locally rather than their cause.Had the PET experiments shown only unilateral activity in(Werhahnet al., 1994), the demonstration of an rCBF increase

in M1 does not give a direct measure of the activity of these M1 in XKS subjects, they would have been decisive. However,given the finding of bilateral activity, the PET data do notoutput cells.

1228 M. Kramset al.

Kallmann FJ, Schoenfeld WA, Barrera SE. The genetic aspects offinally resolve the issue. Nonetheless, we have suggested thatprimary eunuchoidism. Am J Ment Deficiency 1944; 48: 203–36.the PET findings could be consistent with theLemon RN. Functional properties of monkey motor cortex neuroneselectrophysiological findings of Maystonet al. (1997) if it isreceiving afferent input from the hand and fingers. J Physiol (Lond)supposed that the activation of the motor cortex opposite the1981; 311: 497–519.mirroring hand is the result of the movement rather than itsMatelli M, Luppino G. Cortical projections of motor thalamus. In:cause. The issue could be finally resolved if there wereMinciacchi D, Molinari M, Macchi G, Jones EG, editors. Thalamicstructural evidence on the development of the pyramidal tractnetworks for relay and modulation. Oxford: Pergamon Press, 1993:in XKS subjects. We have such studies in progress.165–74.

Mayer M, Botzel K, Paulus W, Plendl H, Prockl D, Danek A.Movement-related cortical potentials in persistent mirror movements.Electroencephalogr Clin Neurophysiol 1995; 95: 350–8.

ReferencesAndersson G. Cortico-cortical mediation of short-latency (lemniscal)

Mayston MJ, Harrison LM, Quinton R, ıStephens JA, Kramm M,sensory input to the motor cortex in deeply pentobarbitoneBouloux P-MG. Mirror movements in X-linked Kallmann’sanaesthetized cats. Acta Physiol Scand 1995; 153: 381–92.syndrome: I. A neurophysiological study. Brain 1997; 120: 000–000.

Bernard RA, Goran DA, Nordell BA, Cooper TG, Conlon TG, NerosMuakkassa KF, Strick PL. Frontal lobe inputs to primate motor cortex:C, et al. Imaging the motor and sensory cortical areas involved in theevidence for four somatotopically organized ‘premotor’ areas. Brainperformance of a simple task: a functional MRI study. Proc Int SocRes 1979; 177: 176–82.Magn Reson Med 1996; 2: 1863.

Oldfield RC. The assessment and analysis of handedness: theBrinkman C. Supplementary motor area of the monkey’s cerebralEdinburgh Inventory. Neuropsychologia 1971; 9: 97–113.cortex: short- and long-term deficits after unilateral ablation and the

effects of subsequent callosal section. J Neurosci 1984; 4: 918–29.Quinton R, Duke V-M, de Zoysa PA, Bouloux P-MG. Theneurobiology of Kallmann’s syndrome. Hum Reprod 1996a; 11 NatlBrittonTC, MeyerBU, BeneckeR. Centralmotor pathways in patientsSuppl: 121–7.with mirror movements [published erratum appears in J Neurol

Neurosurg Psychiatry 1991; 54: 510]. J Neurol Neurosurg PsychiatryQuinton R, Duke V-M, de Zoysa PA, Platts AD, Valentine A, Kendall1991; 54: 505–10. B, et al. The neuroradiology of Kallmann’s syndrome: a genotypic

and phenotypic analysis. J Clin Endocrinol Metab 1996b; 81: 3010–7.Cheney PD, Fetz EE. Corticomotoneuronal cells contribute to long-

Rao SM, Binder JR, Bandettini PA, Hammeke TA, Yetkin FZ,latency stretch reflexes in the rhesus monkey. J Physiol (Lond) 1984;Jesmanowicz A, et al. Functional magnetic resonance imaging of349: 249–72.complex human movements. Neurology 1993; 43: 2311–8.

Cohen LG, Meer J, Tarkka I, Bierner S, Leiderman DB, DubinskyRobb RA, Hanson DP. A software system for interactive andRM, et al. Congenital mirror movements. Abnormal organization ofquantitative visualization of multidimensional biomedical images.motor pathways in two patients. Brain 1991; 114: 381–403.Australas Phys Eng Sci Med 1991; 14: 9–30.

Conrad B, Kriebel J, Hetzel WD. Hereditary bimanual synkinesisRouiller EM, Babalian A, Kazennikov O, Moret V, Yu XH,combined with hypogonadotropic hypogonadism and anosmia in fourWiesendanger M. Transcallosal connections of the distal forelimbbrothers. J Neurol 1978; 218: 263–74.representations of the primary and supplementary motor cortical areas

Danek A, Heye B, Schroedter R. Cortically evoked motor responsesin macaque monkeys. Exp Brain Res 1994; 102: 227–43.in patients with Xp22.3-linked Kallmann’s syndrome and in female

Shibasaki H, Nagae K. Mirror movement: application of movement-gene carriers. Ann Neurol 1992; 31: 299–304.related cortical potentials. Ann Neurol 1984; 15: 299–302.

Evans AC, Marrett S, Torrescorzo J, Ku S, Collins L. MRI-PET Stephan KM, Fink GR, Passingham RE, Silbersweig D, Ceballoscorrelation in threedimensions usinga volume-of-interest (VOI) atlas.Baumann AO, Frith CD, et al. Functional anatomy of the mentalJ Cereb Blood Flow Metab 1991; 11: A69–78. representation of upper extremity movements in healthy subjects. J

Neurophysiol 1995; 73: 373–86.Evans AC, Collins DL, Mills SR, Brown ED, Kelly RL, Peters TM.3D statistical neuroanatomical models from 305 MRI volumes. ProcTalairach J, Tournoux P. Co-planar stereotaxic atlas of the humanIEEE Nucl Sci Symp Med Imag Conf 1993: 1813–7. brain. Stuttgart: G. Thieme, 1988.Forget R, Boghen D, Attig E, Lamarre Y. Electromyographic studiesvan der Linden C, Bruggeman R. Bilateral small-hand-muscle motorof congenital mirror movements. Neurology 1986; 36: 1316–22. evoked responses in a patient with congenital mirror movements.

Electromyogr Clin Neurophysiol 1991; 31: 361–4.Friston KJ, Ashburner J, Frith CD, Poline JB, Heather JD, FrackowiakRSJ. Spatial registration and normalization of images. Hum BrainWerhahn KJ, Fong JK, Meyer BU, Priori A, Rothwell JC, Day BL,Mapp 1995a; 3: 165–89. et al. The effect of magnetic coil orientation on the latency of surface

EMG and single motor unit responses in the first dorsal interosseousFriston KJ, Holmes AP, Worsley KJ, Poline JB, Frith CD, Frackowiakmuscle. Electroencephalogr Clin Neurophysiol 1994; 93: 138–46.RSJ. Statistical parametric maps in functional brain imaging: A

general linear approach. Hum Brain Mapp 1995b; 2: 189–210. Zeffiro TA, Hallett M. Regional cerebral blood flow changes duringactive and passive finger movements: a PET study [abstract]. Soc

Ghosh S, Brinkman C, Porter R. A quantitative study of theNeurosci Abstr 1992; 18: 938.distribution of neurons projecting to the precentral motor cortex inthe monkey (M. fascicularis). J Comp Neurol 1987; 259: 424–44.

Jueptner M, Weiller C. Does measurement of regional cerebral bloodReceived November 21, 1996. Accepted February 10, 1997flow reflect synaptic activity. [Review]. Neuroimage 1995; 2: 148–56.