Effects of subthalamic nucleus stimulation on motor cortex excitability in Parkinson’s disease

J Neurol (2002) 249 : 1689–1698DOI 10.1007/s00415-002-0906-y ORIGINAL COMMUNICATION

Stéphane ThoboisPeter DomineyValérie FraixPatrick MertensMarc GuenotLuc ZimmerPierre PollakAlim-Louis BenabidEmmanuel Broussolle

Effects of subthalamic nucleus stimulationon actual and imagined movement in Parkinson’s disease : a PET study

Abbreviations

PET: Positron Emission TomographyPD: Parkinson’s diseaseSMA: Supplementary Motor AreaDLPFC: Dorsolateral Prefrontal Cortex

STN: Subthalamic NucleusrCBF: regional Cerebral Blood FlowGPi: Internal Globus PallidusBA: Brodmann area

JON

906

Received: 7 March 2002Received in revised form: 6 June 2002Accepted: 12 June 2002

Dr. S. Thobois (!) · E. Broussolle, MD, PhDService de Neurologie D (Pr. G. Chazot)Hôpital Neurologique Pierre Wertheimer59 Bd Pinel69003 Lyon, FranceTel.: +33-4 72/35 72 18Fax: +33-4 72/35 73 51E-Mail : [email protected]

S. Thobois, MD · L. Zimmer, PharmD, PhD ·E. Broussolle, MD, PhDCERMEP, Cyclotron UnitThe Neurological Hospital PierreWertheimerLyon, France

P. Dominey, PhDInstitute of Cognitive Science, CNRSLyon, France

V. Fraix, MD · P. Pollak, MD · A.-L. Benabid, MD, PhDDepartments of Neurology and Neurosurgeryand INSERM U318CHU of GrenobleGrenoble, France

P. Mertens, MD, PhD · M. Guenot, MDDepartment of Neurosurgery A(Pr Sindou)Neurological Hospital Pierre WertheimerLyon, France

! Abstract Background PET stud-ies in moderately affected Parkin-son’s disease (PD) patients revealabnormal cerebral activation dur-ing motor execution and imagery,but the effects of subthalamic nu-cleus (STN) stimulation are notwell established. Objectives to as-sess the effect of STN stimulationon cerebral activation during ac-tual and imagined movement inpatients with advanced PD. Meth-ods seven severely affected PD pa-tients treated with bilateral STNstimulation were studied with PETand H2

15O. The following condi-tions were investigated: (1) rest; (2)motor execution of a sequentialpredefined joystick movement withthe right hand and (3) motor im-agery of the same task. Patientswere studied with and without leftSTN stimulation while right stimu-lator remained off. Results WithoutSTN stimulation, the primary mo-tor cortex was activated only dur-ing motor execution whereas thedorsolateral prefrontal cortex(DLPFC) was activated only duringmotor imagery. An activation of the

supplementary motor area (SMA)was seen during both motor execu-tion and motor imagery. Left STNstimulation during motor execu-tion increased the regional cerebralblood flow (rCBF) bilaterally in theprefrontal cortex including DLPFC,in the left thalamus and putamen.In addition, a reduction of rCBFwas noted in the right primary mo-tor cortex, inferior parietal lobeand SMA. Under left STN stimula-tion, during motor imagery, rCBFincreased bilaterally in the DLPFCand in the left thalamus and puta-men and decreased in the left SMAand primary motor cortex. Conclu-sion STN stimulation during bothmotor execution and imagerytends to improve the functioning ofthe frontal-striatal-thalamic path-way and to reduce the recruitmentof compensatory motor circuitsnotably in motor, premotor andparietal cortical areas.

! Key words PET · Parkinson’sdisease · rCBF · motor execution ·motor imagery · subthalamicnucleus · deep brain stimulation

1690

Introduction

Parkinson’s disease (PD) is characterized by the associ-ation of tremor, akinesia and rigidity related to adopaminergic deficiency of the nigrostriatal pathway.This leads to a dysfunction of the cortical-basal ganglialoops resulting in a profound inhibition of the thalamusand subsequently of the frontal cortex, especially thesupplementary motor area (SMA) and the dorsolateralprefrontal cortex (DLPFC), areas which are particularlyimplicated in selection and preparation of movement [1,12, 30]. Positron Emission Tomography (PET) and func-tional magnetic resonance imaging (fMRI) studies dur-ing execution of a manual motor task in PD have shown,on the one hand, an hypoactivation (i. e. a decreased re-gional cerebral blood flow (rCBF)) in the putamen, cin-gulate cortex and DLPFC and, on the other hand, a com-pensatory hyperactivation (i. e. an increased rCBF) inthe primary motor, lateral premotor and parietal cortexand in the cerebellum [24, 33–35, 37, 38, 44]. Concerningthe SMA, several rCBF studies have shown an hypoacti-vation of this area compared with controls during a mo-tor task [24,33–35,38,44].This effect may be reversed bydopaminergic medical treatment [21, 24, 34]. However, arecent PET study demonstrated that SMA recruitmentremains possible with complex motor sequences com-pared with a simple motor task [6]. Apomorphine chal-lenge or stereotactic surgical procedures achieve theirefficacy in alleviating parkinsonian symptoms by re-ducing the inhibitory output of these structures on thethalamocortical pathway as suggested by electrophysio-logical recordings [3, 22, 25, 26, 41]. This is supported byresults from recent PET studies during actual executionof a motor task showing that hypoactivation of the ros-tral SMA and of the DLPFC may be corrected by sub-thalamic nucleus (STN) stimulation or by pallidotomy[7, 17, 19, 27, 32, 39].

A complementary approach to studying motor pro-cessing independant of execution relies on the motorimagery paradigm, which consists of the internal re-hearsal of a given movement without overt motor out-put. A number of behavioral, electrophysiological, and

functional imagery experiments suggest that motor ex-ecution and motor imagery share common neural mo-tor circuits [10, 11, 36, 42, 47, 48]. Accordingly, chrono-metric, electrophysiological and PET studies providestrong arguments showing that motor imagery is alsoaffected in PD [8, 9, 14, 40, 45].

The ability of STN stimulation to modify the abnor-mal activation pattern during a motor task in frontalcortical and basal ganglia regions in PD is not well es-tablished. In addition, no PET activation study, to ourknowledge, has analysed the effects of STN or GPi deepbrain stimulation or pallidotomy on motor imagery.This approach would be of great interest since motorimagery relies on an internal representation of actionand permits the study of this aspect of motor prepara-tion.We therefore decided to carry out a PET rCBF studyusing H2

15O to examine in PD patients the modulationof cerebral activation induced by STN stimulation dur-ing motor execution and motor imagery. Our hypothe-sis was that STN stimulation could restore the frontal-striatal-thalamic pathway activation on one hand, andreduce the recruitment of accessory motor pathway ar-eas on the other hand, during both actual and imaginedmovement.

Material and methods

! Subjects

Seven severely affected PD patients were studied (mean age: 56.3± 11.4 years, range: 49–65; six men, one woman). The patients fulfilledthe UK Parkinson’s Disease Brain Bank criteria for idiopathic Parkin-son’s disease [18]. All the patients had preserved Dopa-effectiveness,severe Off states,motor fluctuations and dyskinesias.PD patients weretreated for at least three months prior to the study by chronic electri-cal stimulation of the STN (monopolar stimulation; mean ± SD: am-plitude: 3.16 ± 0.4 V; pulse width: 60 µs; frequency: 134.3 ± 7 Hz). Thesurgical procedure was performed as previously described [28]. Theimplantation of electrodes (Model 3389; Medtronic, Minneapolis,MN) was done under stereotactic guidance and the electrodes werethen connected to a telemetrically controllable pulse generator (ItrellII or Kinetra, Medtronic). Four patients were operated on in Grenobleand three in Lyon. All patients had bilateral akinetic-rigid signs with-out tremor. The patients’ clinical data are presented in Table 1. The

Patient Disease duration UPDRS motor score Dopaminergic Treatmentn°/Sex/Age (years) (Off drug) (Dopa equivalent/mg)

On stimulation Off stimulation

1/M/60 27 14.5 36.5 100 (B)2/M/52 17 27 68 300 (D, B)3/M/65 17 14.5 29.5 450 (D, R)4/M53 13 15 56 700 (D, R)5/M/54 11 5 39 06/F/49 15 9 37 07/M/61 17 20 31 900 (D, R)mean ± SD 16.7±5 15.2±8.5 46.2±15 350±350

Dopa equivalent: 100 mg L-dopa (D) = 10 mg Bromocriptine (B) = 5 mg Ropinirole (R)

Table 1 Clinical characteristics of the patients at thetime of PET study

1691

seven patients were right-handed as assessed by the Annett handed-ness inventory [2]. UPDRS motor score was determined off medica-tion with and without stimulation at the time of the PETs [15]. Sub-jects were familiarized and trained on motor imagery tasks by usinga questionnaire and exercises given a few weeks before the PET study[20]. All patients were off-drug at the time of PET study and the lastdose of antiparkinsonian treatment was taken at least 12 hours beforethe scan acquisition.

The present study was performed after approval by the Lyon Uni-versity Hospitals Ethics Committee.All subjects participated after thenature of the procedure had been fully explained and signed an in-formed consent form according to the declaration of Helsinki.

! Activation tasks

Subjects were scanned while imagining or executing a predefined se-quential motor task with the right hand. Subjects lay down on the bedof the PET scanner with their eyes closed. The left stimulator wasswitched on or off 20 minutes before the PET scanning. The followingsix conditions were studied: 1) rest without stimulation (RS–); 2) restwith effective unilateral left-sided stimulation (RS+); 3) execution ofthe motor task without stimulation (ES–); 4) execution of the motortask with effective unilateral left-sided stimulation (ES+); 5) imagi-nation of the motor task without stimulation (IS–); 6) Imagination ofthe motor task with effective unilateral left-sided stimulation (IS+).The efficacy of left stimulation was assessed by testing akinesia withthe 23rd item of UPDRS motor score. This task consisted in repetitiveright index to thumb opposition movements and was quoted on a 4points score (0: normal; 4:movement impossible) (mean score stimu-lation Off: 2.28 ± 0.5 versus mean score stimulation On 0.9 ± 0.7;p < 0.01). The right stimulator was off throughout the study. The con-ditions were pooled according to the stimulation state in 2 blocks of3 conditions each in duplicate: 1) On stimulation (IS + , ES + , RS +);2) Off stimulation (IS–, ES–, RS–). The order of the blocks was coun-terbalanced from one patient to the next. Each of the motor condi-tions was repeated once in a variable order for a total of 12 sessions.Prior to each scanning session, and after general instructions hadbeen given, a few practice trials for each condition were performed toensure that the task was properly understood. The motor task con-sisted in moving a joystick with the right hand in three sequential andpredefined directions. This task was performed after an oral Go sig-nal was given by the examiner and was repeated in the same mannerduring 90 seconds for each condition, as fast and accurately as possi-ble. The end of each session was indicated by an oral Stop signal.

To control for the subject’s motor imagery performances, weasked the subjects, the day prior to PET-scanning, to execute or imag-ine, four times,on and off stimulation, the motor task described aboveand we measured the mean duration needed to perform these tasks.This was to ensure that motor imagery completion time correlatedwith motor execution completion time which indirectly reflects thecorrect imagination of the motor task. For this chronometric study,patients were off medication for 12 hours and the right stimulator wasswitched off 10 minutes before starting the experiment. The neces-sary checking of the motor imagery paradigm also explains why wedecided to use in our experimental procedure a predefined repetitivesequential motor task rather than movement in freely selected direc-tions.

! Scanning procedure

The head of the subject was maintained in a fixed position using athermoformed mask. Checking of the head position throughout theexamination was made by laser alignment along with reference pointson Reid’s line before and after each session. PET scans were acquiredusing a Siemens CTI HR + tomograph (CTI/Siemens, Knoxville, TN)in a 3 dimensional mode. Transmission data were acquired using ro-tating sources filled with 68Ge/68Ga. Images were reconstructed by 3D

filtered back projection (Hanning filter; cut-off frequency, 0.5 cy-cles/pixel), giving a transaxial resolution of 6.5 mm full width at halfmaximum, and displayed in a 128 x 128 pixel format with 63 planescreating approximately 2-mm cubic voxels. rCBF was estimated byrecording the distribution of radioactivity following an intravenousinjection of 333 MBq of H2

15O through a forearm cannula placed intothe brachial vein. The integrated counts were collected for 90 seconds,starting 20 seconds after the injection. For data analysis we only con-sidered the 60 seconds corresponding to the maximum radioactivity.A 10-minute interval was necessary between each test condition foradequate radioactivity decay.

! PET data and statistical analysis

Image and statistical analysis were performed in MATLAB 5.3 (MathWorks, Natick, Mass., USA) using software for statistical parametricmapping (SPM 99, Wellcome Department of Cognitive Neurology,MRC Cyclotron Unit, London, UK). Briefly, series of scans werealigned to the first scan with an automated algorithm for head move-ments correction and then normalized into a standard stereotacticspace [43]. The images were smoothed with an isotropic Gaussiankernel (full width half maximum 14 mm for all directions) to allow forinterindividual gyral variation and to improve the signal: noise ratio.Global differences in rCBF were covaried out for all voxels and com-parisons across conditions were made using t statistics with appro-priate linear contrasts,and then converted to Z-scores.Coordinates ofthe activated foci obtained from SPM 99 were then converted in Ta-lairach coordinates using the appropriate formula (to transformMcGill-SPM96-coordinates into Talairach 88-SPM 95 coordinates:x’ = 0.88x + 0.8; y’ = 0.97y – 3.32; z’ = 0.05y + 0.88z – 0.44). Only re-gional activation significant at p < 0.001 at each pixel, uncorrected formultiple comparisons (Z > 3.10), were retained. Uncorrected p valuewas accepted because of an a priori strong hypothesis on the SMA,prefrontal cortex and thalamic regions. The following comparisonswere made: 1) STN stimulation-induced rCBF changes during motorexecution. Increased activation induced by STN stimulation duringmotor execution was analysed by the following contrast : (ES+) –(ES–). Decreased activation induced by STN stimulation during mo-tor execution was analysed by the following contrast : (ES–) – (ES+);2) STN stimulation-induced rCBF changes during motor imagery. In-creased activation induced by STN stimulation during motor imagerywas analysed by the following contrast : (IS+) – (IS–). Decreased acti-vation induced by STN stimulation during motor imagery wasanalysed by the following contrast : (IS–) – (IS+); 3) The interactionbetween motor task and STN stimulation for motor execution andimagery by the following contrasts: [(ES+) – (RS+)] – [(ES–) – (RS–)]and [(IS+) – (RS+)] – [(IS–) – (RS–)] respectively. 4) The activationprofile during motor imagery and motor execution versus rest with-out stimulation: (IS–) – (RS–) and (ES–) – (RS–) respectively.

Results

! Task performance in PD patients

As mentioned in the methods section, these data weremeasured the day prior to PET scanning, subjects beingoff drug for at least 12 hours.Sequence completion timesof the right hand for imagery and execution, with andwithout stimulation were analysed in a multi-factor re-peated measures ANOVA in which the within-subjectsvariables were Task (execution or imagery), Stimulation(On or Off), Epoch (1st and 2nd) and Repeat (4 condi-tions). The dependent measure was the completion time

1692

for a given trial. There was no significant difference forimagery vs. execution completion time (F(1.4) = 2.0,p = 0.23). In addition, completion times were reducedfor stimulation On (4.74 s) versus stimulation Off (5.76s) sequences but this effect did not reach significance (F[1, 4]= 6.13, p = 0.068). This reveals that the beneficial ef-fects of stimulation were present for execution and mo-tor imagery. Furthermore, both with and without stim-ulation, motor imagery completion times werecorrelated with motor execution completion times (mo-tor imagery completion time = 1.7 + 0.86 x motor execu-tion completion time; r = 0.68, p = 0.03).

! Main effects of motor execution without stimulation(Table 2)

When motor execution of the right hand was comparedwith rest without stimulation, significant activation fociwere disclosed in the left primary motor cortex, SMAand in the right cerebellum, inferior parietal lobe (BA40), inferior frontal gyrus (BA 45) and lateral premotorcortex (BA 6). No significant activation of the basal gan-glia was noted.

! Main effects of motor imagery without stimulation(Table 2; Fig 1)

When motor imagery of the right hand was comparedwith rest without stimulation, significant activationswere seen bilaterally in the SMA, DLPFC (BA 9) and in-

ferior parietal lobe (BA 40), and ipsilaterally in the rightsuperior parietal lobe (BA 7) and the right cerebellarhemisphere.

! STN Stimulation-induced rCBF changes during motorexecution (Table 3, Fig 2A and 2B)

Left STN stimulation during motor execution with theright hand produced an increased left prefrontal (BA 11,9 and 45), thalamic and putaminal activation. In addi-tion, the interaction between the motor task and thestimulation revealed a specific activation located in theright DLPFC. In contrast, a reduction of rCBF was dis-closed in the right primary motor cortex, SMA, and in-ferior parietal lobe (BA 40).

! STN Stimulation-induced rCBF changes during motorimagery (Table 3, Fig 3A and B)

Left STN stimulation during motor imagery with theright hand was associated with an increased activationbilaterally in the prefrontal cortex including DLPFC andthe left thalamus and putamen. The study of the inter-action between motor imagery and the stimulationshowed, as during motor execution, a specific activationlocated in the right DLPFC. In contrast, reduction ofrCBF was disclosed in the left SMA and primary motorcortex.

A) Motor execution vs rest

Areas Side (L/R) Talairach coordinates Z score p uncorr

x y z

Primary motor cortex L –24 –25 49 6.61 0.0001SMA L –8 –2 56 5.23 0.0001Lateral premotor cortex R 20 –1 57 5.42 0.0001Inferior parietal lobe (BA 40) R 45 –41 41 4.52 0.0001Inferior frontal gyrus (BA 46) R 34 28 3 3.65 0.0001Cerebellum R 22 –43 –24 5.85 0.0001

B) Motor imagery vs rest

Areas Side (L/R) Talairach coordinates Z score p uncorr

x y z

Prefrontal cortex (DLPFC) L –47 9 35 4.75 0.0001R 29 46 28 3.39 0.0001

SMA L –3 1 57 5.45 0.0001R 11 1 64 5.43 0.0001

Superior parietal lobe (BA 7) R 1 –63 58 3.60 0.0001Inferior parietal lobe (BA 40) L –41 –35 24 4.44 0.0001

R 55 –41 27 4.30 0.0001Cerebellum R 29 –43 –28 3.66 0.0001

Table 2 Sites of activation during motor tasks vsrest (Off stimulation)

1693

Discussion

The present study reveals two important findings. First,STN stimulation during motor imagery and executioninduces increased rCBF in the prefrontal cortex includ-ing DLPFC and restores a more physiological frontal-

striatal-thalamic activation. Second, STN stimulationduring motor execution and imagery is associated witha decreased rCBF in accessory motor pathways notablyin the motor, premotor and parietal areas.

Fig. 1 Activated areas during motor imagery versusrest without stimulation. The statistical parametricmaps are displayed in the anatomical space of Ta-lairach and Tournoux as a maximum intensity projec-tion superimposed onto lateral (left and right hemi-spheres), anterior and horizontal views of the singlesubject brain MRI of SPM99.

A) Motor execution

Areas Side (L/R) Talairach coordinates Z score p uncorr

x y z

Enhanced activationPrefrontal cortex (BA 11) L –22 42 –11 3.11 0.001Prefrontal cortex (BA 9)/DLPFC R 17 63 31 3.10 0.001Inferior frontal gyrus (BA 45) L –31 28 6 3.15 0.001Thalamus L –11 –12 0 4.14 0.0001Putamen L –22 13 0 3.21 0.001

Reduced activationSMA R 4 26 52 4.21 0.0001Primary motor cortex (BA 4) R 4 –39 59 3.22 0.001Inferior parietal lobe (BA 40) R 38 –64 42 3.12 0.001

B) Motor imagery

Areas Side (L/R) Talairach coordinates Z score p uncorr

x y z

Enhanced activationPrefrontal cortex (BA 10)/DLPFC R 18 67 24 4.07 0.0001

L –11 71 19 3.5 0.0001Thalamus L –10 –10 –1 3.85 0.0001Putamen L –18 19 –5 3.44 0.0001

Reduced activationSMA L –3 –8 62 3.61 0.0001Primary motor cortex L –8 –24 70 4.42 0.0001

Table 3 Modifications of cerebral activation in-duced by STN stimulation

1694

! Execution of actual movement with and without STN stimulation

Activation profile during motor execution without stimulation

The activation pattern during motor execution with theright hand compared with rest without stimulationessentially concerns the contralateral primary motorcortex, caudal SMA, ipsilateral cerebellum and the left DLPFC. The rostral SMA is not activated which is in agreement with previous studies performed in PD.In our study, the absence of a control group does notallow us to draw conclusions on an eventual reductionof rostral SMA rCBF. Furthermore, it must be noticed

that the classical hypoactivation of the anterior SMAnoted in earlier studies in PD [24, 33] is now morediscussed and clearly depends on the motor task.Indeed, the more the movement is complex and self-initiated, the more the rostral SMA rCBF is reduced inPD compared with controls [6, 23]. In addition, wefound a preserved activation of the caudal SMA whichis in accordance with a recent fMRI study which evenmore clearly demonstrated in complex sequentialmovement an hyperactivation of the caudal SMA in PDpatients compared with controls [37]. Similar result hasbeen demonstrated by PET [40]. Since the caudal SMAis considered by some authors to belong to a fron-toparietal pathway, it may be normally activated oroveractivated in PD similar to the parietal, premotor

Fig. 2 (a) Relative increase of activation induced byleft STN stimulation during motor execution. The sta-tistical parametric maps are displayed in theanatomic space of Talairach and Tournoux as a maxi-mum intensity projection superimposed onto lateral(left and right hemispheres), anterior and internalviews of the single subject brain MRI of SPM99. Thesignificant overactivations concern the left thalamusand putamen and bilaterally the prefrontal cortex in-cluding DLPFC. Non-significant activations appearalso in the left ventral prefrontal BA 10.

(b) Relative decrease of activation induced by left STNstimulation during motor execution. The statisticalparametric maps are displayed in the anatomic spaceof Talairach and Tournoux as a maximum intensityprojection superimposed onto lateral (right hemi-sphere), anterior, internal and top views of the singlesubject brain MRI of SPM99. The reduction of activa-tion concerns the right SMA, inferior parietal lobe andprimary motor cortex.

Internal view

1695

and motor cortex, rather than being underactivated likethe rostral SMA [23, 29, 37, 38, 40].

Effects of STN stimulation on motor execution activationprofile

An increased activation ipsilateral to the stimulation isdisplayed in the left thalamus, putamen and bilaterallyin the prefrontal cortex notably the DLPFC. This findingunderlines the restoration of a more physiologicalfrontal-striatal-thalamic circuit by STN stimulation al-though the precise mechanims of such stimulation re-mains debated [4]. This effect may be explained by animproved functioning of the thalamocortical excitatorypathway and is in accordance with recent studies [7, 27].

Similar results concerning the prefrontal cortex havealso been reported after pallidotomy but not after GPistimulation in particular for the DLPFC [27, 39]. Inter-estingly the overactivation observed in the right DLPFC(i. e. controlateral to the stimulation and ispilateral tothe movement) has also been found by Ceballos-Bau-mann et al. [7]. In a recent PET study, the right DLPFC,and not the left one, appeared hypoactivated, comparedwith controls, while PD patients performed a joystickmotor task in freely selected directions with the righthand [40].

Several authors have reported an increased rCBF inthe SMA following STN or GPi stimulation, pallidotomyor apomorphine challenge [7, 17, 19, 24, 27, 39]. In con-trast, in our work, STN stimulation induces a reduction

Fig. 3 (a) Relative increase of activation induced byleft STN stimulation during motor imagery. The sta-tistical parametric maps are displayed in theanatomic space of Talairach and Tournoux as a maxi-mum intensity projection superimposed onto lateral(right hemisphere), anterior, internal and top viewsof the single subject brain MRI of SPM99. The signifi-cant overactivations concern the left thalamus andputamen and bilaterally the prefrontal cortex notablythe DLPFC. Non-significant activation appears also inthe pons and anterior cingulate.

(b) Relative decrease of activation induced by left STNstimulation during motor imagery. The statisticalparametric maps are displayed in the anatomic spaceof Talairach and Tournoux as a maximum intensityprojection superimposed onto lateral (left hemi-sphere), anterior, sagittal and top views of the singlesubject brain MRI of SPM99. The reduction of activa-tion concerns the left SMA and primary motor cortex.

1696

of SMA activation. This result may be interpreted in thecontext of a global reduction of cortical recruitment in-duced by the stimulation concomitantly with a greatmotor improvement. This may indicate that STN stimu-lation tends to suppress the compensatory mechanismsdeveloped in PD to improve motor performances [35,37,38]. Indeed,a reduction of rCBF is also noted in the rightinferior parietal lobe, primary motor cortex when thestimulation was on compared with the off condition.Preliminary results from another group are in agree-ment with our findings, with a decreased activation ofcaudal SMA which may be part of these accessory motorpathways, and primary motor cortex induced by STNstimulation [31].

! Imagined movement with and without STN stimulation

Activation profile during motor imagery without stimulation

A bilateral activation of the SMA and DLPFC is dis-played in our study during motor imagery with the righthand. However, no SMA or DLPFC hypoactivations canbe assessed based on our data as no control group wasstudied. Three previous PET experiments including onefrom our group investigated the consequences ofparkinsonian motor disability on motor imagery activa-tion profile [9, 40, 45]. They revealed a preserved activa-tion of the SMA in PD, which is reinforced by the presentstudy. Our data overlap with those reported by Samuelet al. [40] although we observe a bilateral instead of leftDLPFC activation and no lateral premotor cortex acti-vation. In our preceding study, compensatory overacti-vations were noted in the primary motor cortex andsuperior parietal lobe during motor imagery in mid-stage PD [45]. In addition, SMA was activated contralat-

erally to the motor task while the DLPFC was activatedipsilaterally to the motor task. Likewise, Cunningtonet al. [9] noted a preserved activation of the SMA and in-ferior parietal lobe during motor imagery on but also offantiparkinsonian treatment, compared with normalsubjects. Taken together all these findings differ fromthe SMA hypoactivation previously reported duringmotor execution in PD and emphasize the role of DLPFCactivation during motor imagery [24].

Effects of STN stimulation on motor imagery activation profile

STN stimulation induced a significant increase of rCBFlocated in the left thalamus, putamen and bilaterally inthe DLPFC. This pattern is very close to the one ob-served during motor execution. This suggests that mo-tor execution and motor imagery share common func-tional circuits as already demonstrated in normalsubjects and PD patients [11, 14]. Our results tend toprove that STN stimulation could improve motor prepa-ration, decision making and not only execution of theactual movement.

The other important finding is that STN stimulationreduces SMA activation during motor imagery. Recentworks demonstrate that the early component of theBereitschaftpotential (BP) which is usually reduced inamplitude in PD, does not change during STN or GPistimulation [5, 13, 46]. Our finding appears important asSMA plays a major role in the generation of this earlycomponent of the BP, and argues that STN stimulationcould have little effect on SMA activation during motorpreparation and consequently motor imagery [46].

! Acknowledgements This work was supported by a grant from theAssociation France Parkinson. The authors thank F. Lavenne, C.Pierre, D. Le Bars and the nurses of the CERMEP for technical supportand J. Decety and N. Costes for helpful discussions.

References

1. Alexander GE, Crutcher MD, De LongMR (1990) Basal ganglia thalamocorti-cal circuits: parallel substrates for mo-tor, oculomotor, prefrontal and limbicfunctions. (Review) Prog Brain Res85:119–146

2. Annett M (1970) A classification ofhand preference by association analy-sis. Br J Psychol 61:303–321

3. Benazzouz A, Piallat B, Pollak P, Ben-abid AL (1995) Responses of substan-tia nigra pars reticulata and globuspallidus complex to high frequencystimulation of the subthalamic nucleusin rats: electrophysiological data. Neu-rosci Lett 189:77–80

4. Benazzouz A, Hallett M (2000) Mecha-nism of action of deep brain stimula-tion. Neurology 55 (Suppl 6):S13–S16

5. Brown RG, Limousin Dowsey P, BrownP, Jahanshahi M, Pollak P, Benabid AL,Rodriguez-Oroz MC, Obeso J, RothwellJC (1999) Impact of deep brain stimu-lation on upper limb akinesia inParkinson’s disease. Ann Neurol45:473–488

6. Catalan MJ, Ishii K, Honda M, Samii A,Hallett M (1999) A PET study of se-quential finger movements of varyinglength in patients with Parkinson’s dis-ease. Brain 122:483–495

7. Ceballos-Bauman AO, Boecker H,Bartenstein P, von Falkenhayn I, Rie-scher H, Conrad B, Moringlane JR,Alesch F (1999) A positron emissiontomographic study of subthalamic nu-cleus stimulation in Parkinson’s dis-ease. Arch Neurol 56:997–1003

8. Cunnington R, Iansek R, Johnson KA,Bradshaw JL (1997) Movement-relatedpotentials in Parkinson’s disease. Mo-tor imagery and movement prepara-tion. Brain 120:1339–1353

9. Cunnington R, Hughes AJ, O’SullivanJD, Egan GF, Bradshaw JL, Colebatch JG(2001) Motor imagery in Parkinson’sdisease patients: a PET study. Mov Dis-ord 5:849–857

1697

10. Decety J, Perani D, Jeannerod M, Betti-nardi V, Tadary B, Woods R, MazziottaJC, Fazio F (1994) Mapping motor rep-resentations with positron emissiontomography. Nature 371:600–602

11. Decety J, Jeannerod M (1996) Mentallysimulated movements in virtual real-ity: does Fitt’s law hold in motor im-agery? Behav Brain Res 72:127–134

12. De Long MR (1990) Primate models ofmovement disorders of basal gangliaorigin. TINS 13:281–285

13. Dick JP, Rothwell J, Day BL, Cantello R,Buruma O, Gioux M, Beneke R, Be-rardelli A, Thompson PD, Marsden CD(1989) The Bereitschaftpotential is ab-normal in Parkinson’s disease. Brain112:233–244

14. Dominey P, Decety J, Broussolle E,Chazot G, Jeannerod M (1995) Motorimagery of a lateralized sequential taskis asymmetrically slowed in hemi-parkinson’s patients. Neuropsycholo-gia 33:727–741

15. Fahn S, Elton R (1987) Unified Parkin-son’s disease rating scale. In: Fahn S,Marsden CD, Calne D, Goldstein M(eds) Recent developments in Parkin-son’s disease, volume II. FlorhamParks, NJ: MacMillan Healthcare Infor-mation, pp 153–163

16. Filippi MM, Oliveri M, Pasqualetti P,Cicinelli P, Traversa R,Vernieri F,Palmieri MG, Rossini PM (2001) Ef-fects of motor imagery on motor corti-cal output topography in Parkinson’sdisease. Neurology 57:55–61

17. Fukuda M, Mentis M, Ghilardi MF,Dhawan V, Antonini A, Hammerstad J,Lozano AM, Lang A, Lyons K, Koller W,Ghez C, Eidelberg D (2001) Functionalcorrelates of pallidal stimulation forParkinson’s disease. Ann Neurol49:155–164

18. Gibb WRG, Lees AJ (1989) The signifi-cance of the Lewy body in the diagno-sis of idiopathic Parkinson’s disease.Neuropathol Appl Neurobiol 15:27–44

19. Grafton ST, Waters C, Sutton J, Lew MF,Couldwell W (1995) Pallidotomy in-creases activity of motor associationcortex in Parkinson’s disease: apositron emission tomography study.Ann Neurol 37:776–783

20. Hall C, Pongrac J (1983) MovementImagery Questionnaire. University ofWestern Ontario, London, Canada

21. Haslinger B, Erhard P, Kampfe N,Boecker H, Rummeny E, Schwaiger M,Conrad B, Ceballos-Baumann AO(2001) Event-related functional mag-netic resonance imaging in Parkinson’sdisease before and after levodopa.Brain 124:558–570

22. Hutchison WD, Levy R, Dostrovsky JO,Lozano AM, Lang A (1997) Effects ofapomorphine on globus pallidus neu-rons in parkinsonian patients. AnnNeurol 42:767–775

23. Jahanshahi M, Jenkins IH, Brown RG,Marsden CD, Passingham RE, BrooksDJ (1995) Self-initiated versus exter-nally triggered movements. An investi-gation using measurement of regionalcerebral blood flow with PET andmovement-related potentials in nor-mal and Parkinson’s disease subjects.Brain 118:913–933

24. Jenkins IH, Fernandez W, Playford ED,Lees AJ, Frackowiak RSJ, PassinghamRE, Brooks DJ (1992) Impaired activa-tion of the supplementary motor areain Parkinson’s disease is reversed whenakinesia is treated by apomorphine.Ann Neurol 32:749–757

25. Laitinen LV, Bergenheim AT, Hariz MI(1992) Leksell’s posteroventral pallido-tomy in the treatment of Parkinson’sdisease. J Neurosurg 76:53–61

26. Limousin P, Pollak P, Benazzouz A,Hoffmann D, LeBas JF, Broussolle E,Perret JE, Benabid AL (1995) Effect onparkinsonian symptoms of bilateralsubthalamic nucleus stimulation.Lancet 348:91–95

27. Limousin P, Greene J, Pollak P, Roth-well J, Benabid AL, Frackowiak R(1997) Changes in cerebral activitypattern due to subthalamic nucleus orinternal pallidum stimulation inParkinson’s disease. Ann Neurol42:283–291

28. Limousin P, Krack P, Pollak P, Benaz-zouz A, Ardouin C, Hoffmann D, Ben-abid AL (1998) Electrical stimulationof the subthalamic nucleus in ad-vanced Parkinson’s disease. N Eng JMed 339:1105–1111

29. Luppino G, Matelli M, Camarda R, Riz-zolatti G (1993) Corticocortical con-nections of area F3 (SMA-proper) andarea 6 (pre-SMA) in the macaquemonkey. J Comp Neurol 338:114–140

30. Parent A, Sato F, Wu Y, Gauthier J,Lévesque M, Parent M (2000) Organi-zation of the basal ganglia: the impor-tance of axonal collateralization. TINS23 (Suppl):S20–S27

31. Payoux P, Miloudi M, Remy P, DamierP, Samson Y, Rascol O, Agid Y (2000)Changes in cortical motor activity in-duced by subthalamic nucleus stimula-tion in Parkinson’s disease: a PETstudy. Neurology 54 (Suppl 3):P04–049(Abstract)

32. Picard N, Strick PL (1996) Motor areasof the medial wall: a review of their lo-cation and functional activation. CerebCortex 6:342–353

33. Playford ED, Jenkins IH, PassinghamRE, Nutt J, Frackowiak RSJ, Brooks DJ(1992) Impaired mesial frontal andputamen activation in Parkinson’s dis-ease: a PET study. Ann Neurol32:151–161

34. Rascol O, Sabatini U, Chollet F, FabreN, Senard JM, Montastruc JL, Celsis P,Marc-Vergnes JP, Rascol A (1994) Nor-mal activation of the supplementarymotor area in patients with Parkin-son’s disease undergoing long-termtreatment with levodopa. J NeurolNeurosurg Psychiatry 57:567–571

35. Rascol O, Sabatini U, Fabre N, Brefel C,Loubinoux I, Celsis P, Senard JM, Mon-tastruc JL, Chollet F (1997) The ipsilat-eral cerebellar hemisphere is overac-tive during hand movements inakinetic parkinsonian patients. Brain120:103–110

36. Roth M, Decety J, Raybaudi M, Mas-sarelli R, Delon-Martin C, Segebarth C,Morand S, Gemignani A, Décorps M,Jeannerod M (1996) Possible involve-ment of primary motor cortex in men-tally simulated movement: a functionalmagnetic resonance imaging study.Neuroreport 7:1280–1284

37. Sabatini U, Boulanouar K, Fabre N,Martin F, Carel C, Colonnese C, BozzaoL, Berry I, Montastruc JL, Chollet F,Rascol O (2000) Cortical motor reor-ganization in akinetic patients withParkinson’s disease. A functional MRIstudy. Brain 123:394–403

38. Samuel M, Ceballos-Bauman AO, BlinJ, Uema T, Boecker H, Passingham RE,Brooks DJ (1997) Evidence for lateralpremotor and parietal overactivity inParkinson’s disease during sequentialand bimanual movements: a PETstudy. Brain 120:963–976

39. Samuel M, Ceballos-Bauman AO, Tur-janski N, Boecker H, Gorospe A,Linazasoro G, Holmes AP, deLong MR,Vitek JL, Thomas DGT, Quinn NP,Obeso JA, Brooks DJ (1997) Pallido-tomy in Parkinson’s disease increasessupplementary motor area and pre-frontal activation during performanceof volitional movements. Brain120:1301–1313

40. Samuel M, Ceballos-Bauman AO,Boecker H, Brooks DJ (2001) Motorimagery in normal subjects andParkinson’s disease patients: an H2

15OPET study. Neuroreport 26,12(4):821–828

41. Siegfried J, Lippitz B (1994) Bilateralchronic electrostimulation of ventro-posterolateral pallidum: a new thera-peutic approach for alleviating allparkinsonian symptoms. Neuro-surgery 35:1126–1130

42. Stephan KM, Passingham RE, Silber-sweig D, Ceballos-Bauman AO, FrithCD, Frackowiak RSJ (1995) Functionalanatomy of the mental representationof upper extremity movements inhealthy subjects. J Neurophysiol73:373–386

43. Talairach J, Tournoux P (1988) Co-pla-nar stereotaxic atlas of the humanbrain. Stuttgart, Thieme

1698

44. Thobois S, Dominey P, Decety J, PollakP, Gregoire MC, Broussolle E (2000)Overactivation of primary motor cor-tex is asymmetrical in hemiparkinson-ian patients. Neuroreport 11:785–789

45. Thobois S, Dominey P, Decety J, PollakP, Gregoire MC, Le Bars D, Broussolle E(2000) Motor imagery in normal sub-jects and in asymmetrical parkinson’sdisease: a PET study. Neurology55:996–1002

46. Vidailhet M, Stocchi F, Rothwell JC,Thompson PD, Day BL, Brooks DJ,Marsden CD (1993) The bereitschaft-potential preceding simple foot move-ment and initiation of gait in Parkin-son’s disease; Neurology 43:1784–1788

47. Wehner T,Vogt S, Stadler M (1984)Task-specific EMG-characteristics dur-ing mental training. Psychol Res46:389–401

48. Yue G, Cole KJ (1992) Strength in-creases from the motor program: com-parison of training with maximal vol-untary and imagined musclecontractions. J Neurophysiol67:1114–1123