Brain (2001), 124, 2105–2118

Lidocaine and muscimol microinjections insubthalamic nucleus reverse parkinsoniansymptomsRon Levy,1 Anthony E. Lang,3,4 Jonathan O. Dostrovsky,1,3 Peter Pahapill,5 John Romas,1

Jean Saint-Cyr,2,3 William D. Hutchison1,2,3 and Andres M. Lozano2,3

1Department of Physiology, Faculty of Medicine, University Correspondence to: Jonathan O. Dostrovsky, Departmentof Toronto, 2Department of Surgery, University of Toronto, of Physiology Room 3305, Medical Sciences Building,Division of Neurosurgery, 3The Toronto Western Research 1 King’s College Circle, University of Toronto, Toronto,Institute, 4Department of Medicine, University of Toronto, Ontario, Canada M5S 1A8Division of Neurology, The Toronto Western Hospital, E-mail: j.dostrovsky@utoronto.caToronto, Canada and 5Division of Neurosurgery, MedicalCollege of Ohio, Toledo, USA

SummaryInactivation of neurones in the subthalamic nucleus(STN) of the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyri-dine treated monkey model of Parkinson’s disease hasbeen shown to relieve parkinsonian motor symptoms. Inpatients with Parkinson’s disease, neurones in the STNdisplay hyperactive firing rates and rhythmic dischargeactivity such as tremor-related oscillations (3–8 Hz) andsynchronous high-frequency oscillations (15–30 Hz). Inthis study, microinjections of lidocaine (n � 4) andmuscimol, a GABAA receptor agonist (n � 2), wereperformed in the STN of six patients with Parkinson’sdisease to determine whether the focal suppression ofSTN neuronal activity can lead to an improvement intremor, bradykinesia and rigidity. We also report thefirst use of microelectrode recording of the effects ofmicroinjections on neuronal activity in the human brain(n � 2). Microinjections of 10–23 µl of lidocaine producedstriking improvements in bradykinesia, limb tremor andrigidity in three out of three patients. These improvementswere correlated with good therapeutic effects of

Keywords: subthalamic nucleus; Parkinson’s disease; microinjections; muscimol; lidocaine

Abbreviations: DBS � deep brain stimulation; GPe � external segment of the globus pallidus; GPi � internal segment ofthe globus pallidus; MPTP � 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine; STN � subthalamic nucleus; UPDRS � UnifiedParkinson’s Disease Rating Score

IntroductionLocal inactivation of neuronal activity in the subthalamicnucleus (STN) in the 1-methyl-4-phenyl-1,2,3,6-tetra-hydropyridine (MPTP)-treated non-human primate model ofParkinson’s disease has been shown to ameliorateparkinsonian symptoms (Bergman et al., 1990; Wichmann

© Oxford University Press 2001

subsequent STN deep brain stimulation performed in thesame microelectrode trajectories as these injections. Themost dramatic observation following lidocaine injectionswas the appearance of dyskinetic limb movements. In onepatient, simultaneous microelectrode recording during aninjection of 3.5 µl of lidocaine demonstrated a suppressionof neuronal activity at distances of <0.9 mm from theinjection site, but no suppression was observed at ≥1.2 mmfrom the injection site. Microinjections of 5–10 µl ofmuscimol in a region with tremor-related activity resultedin suppression of limb tremor in two out of two patients.Interestingly, in one of these patients, 4 Hz oscillatoryactivity was diminished in a neurone recorded 1.3 mmfrom the injection site, but there was no reduction in themean firing rate or 20 Hz oscillatory activity. These resultsdemonstrate that inactivation of neuronal activity in theSTN of patients with Parkinson’s disease improves motorsymptoms. These findings also suggest that a focal blockof the STN might alter the oscillatory activity of neuroneslocated beyond the inhibited region.

et al., 1994; Guridi et al., 1996). This therapeutic effect isconsistent with the notion of STN hyperactivity or dysfunctionin parkinsonism (Miller and DeLong, 1987; Bergman et al.,1994). Intentional lesioning of the STN in patients withParkinson’s disease has usually been avoided because of the

2106 R. Levy et al.

possible risk of causing severe ballism. However, recentreports have demonstrated good therapeutic effects ofsubthalamotomy in patients with Parkinson’s disease withoutsignificant adverse effects (Gill and Heywood, 1997; Obesoet al., 1997; Muthane et al., 1998; Alvarez et al., 2001;Barlas et al., 2001). Deep brain stimulation (DBS) in theSTN has also been shown to be highly effective in treatingparkinsonian symptoms (Limousin et al., 1995; Kumar et al.,1998; Krack et al., 1998a). To determine the therapeutic andneuronal effects of blocking STN activity in patients withParkinson’s disease, we performed microinjections of thelocal anaesthetic lidocaine and muscimol, a γ-aminobutyricacid-A receptor agonist (GABAA). Lidocaine has the effectof non-selectively blocking axonal fibres of passage as wellas neurones, while muscimol selectively inhibits the cellbodies of neurones.

Elucidating the population behaviour of STN neurones andthe role of the STN in the dynamics of the basal ganglia isessential to extending our understanding of thepathophysiology of Parkinson’s disease (Ryan et al., 1992;Plenz and Kital, 1999; Magill et al., 2000; Levy et al., 2000).The neuronal population in the STN has been shown toexhibit synchronous oscillatory activity (Plenz and Kital,1999; Magill et al., 2000; Brown et al., 2001; Marsden et al.,2001), and it has been suggested that increased oscillatorysynchronization in the basal ganglia underlies thedevelopment of parkinsonian limb tremor (Bergman et al.,1998a; Raz et al., 2000). We recently demonstrated that STNneurones display strongly synchronous 15–30 Hz oscillations(‘high-frequency’ oscillations), in addition to 3–8 Hz tremor-related oscillations, in tremulous patients with Parkinson’sdisease (Levy et al., 2000). Synchronized 3–8 Hz tremor-related activity has also been shown in the internal segmentof the globus pallidus (GPi) (Hurtado et al., 1999) and inthe motor thalamus of parkinsonian patients with Parkinson’sdisease (Levy et al., 1999). Therefore, if parkinsonian limbtremor is due in part to an elevated level of oscillatorysynchronization between basal ganglia subcircuits (Bergmanet al., 1998a), it is conceivable that pharmacological blocks,in addition to inactivating local neuronal activity, mightinfluence the oscillatory behaviour of neurones beyond theregion directly affected by the block. In theory, invasivesurgical therapies could act to reduce limb tremor bydesynchronizing pathological oscillations in the basal ganglia-thalamocortical network (Bergman et al., 1998a; Deuschlet al., 2000). In the present study, we addressed this issueby recording the activity of an STN neurone that displayedboth tremor-related and high-frequency oscillations and waslocated outside a region of pharmacologically blocked tremorcells. This is the first demonstration of the use of simultaneousmicroelectrode recording to assess the effects of micro-injection of substances in the human brain.

MethodsPatient groupStudies of muscimol and lidocaine microinjections into theSTN were performed in seven Parkinson’s disease patients

undergoing microelectrode-guided placement of DBSelectrodes (n � 5) or an STN lesion (n � 2) for the treatmentof the symptoms of Parkinson’s disease. The results for oneof these patients are not presented because postoperativeMRI revealed that microinjections were performed at theventral border of the STN. Results for the rest of the patientsare presented here. This group consisted of two females andfour males who at the time of operation had a mean age of55 years (range 48–67 years). The average duration of thedisease was 12.2 years (�1.8 SE) and all had Parkinson’sdisease for at least 7 years. All patients gave informedconsent and the studies were approved by the HumanExperimentation Committee of the Toronto Hospital.

InjectionsLidocaine (2% Xylocaine, 20 mg/ml without preservative;Astra Pharma Inc., Mississanga, Canada) and muscimol[1 mg/ml (8.8 mM) in sterile saline, 0.2 µm filter sterilized;Sigma, St Louis, Mo., USA] were injected into the STNthrough a stereotactically placed 30 gauge stainless steel tube(Small Parts Inc., Miami Lakes, Fla., USA). The injectioncannula was connected to a 10–15 cm piece of polyethylenetubing (PE 50, inside diameter 0.58 mm) and sealed withepoxy glue. Substances were preloaded in the cannula andpolyethylene tubing and in a 25 µl Hamilton syringe. An airbubble (~1 cm in length) was introduced in the polyethylenetubing to allow the visual determination of the movement ofthe solutions into the STN. The polyethylene tubing wasthen friction fitted over the tip of the Hamilton syringe.Substances were injected at a rate �1.25 µl/min except inone patient in whom lidocaine was injected at 2.5 µl/min(Patient C). The injection cannula was left in the brain for20–25 min after microinjection to reduce the likelihood ofan injected solution diffusing back up the cannula track. Inone case, two injections (saline followed by muscimol) werecarried out by employing two fixed parallel injection cannulae(~350 µm apart) with two separate Hamilton syringes. Thistechnique avoided the need to remove a single injectioncannula in order to insert a second cannula filled with adifferent solution, and thereby reduced the likelihood ofbackflow along the cannula track.

Determination of the target area formicroinjection in STNLocalization of the STN using microelectrode recording isdescribed in detail elsewhere (Hutchison et al., 1998). Briefly,parasagittal trajectories at either 10.5 or 12 mm from themidline passed through the thalamic reticular nucleus and/oranterior thalamus, zona incerta, STN and the substantia nigrapars reticulata (Schaltenbrand and Wahren, 1977). Single unitmicroelectrode recording and stimulation mapping allowedthe identification of physiological landmarks and celllocalization. Exploration of the neuronal activity was carried

STN microinjections in Parkinson’s disease patients 2107

out along the entire dorsal/ventral extent of STN. The maincharacteristics that were identified in order to localize themotor portion of the STN were: neurones with tremor-related activity, neurones that responded to passive or activemovements and microstimulation effects such as tremorreduction or arrest. The anterior–posterior limits of the STNwere delimited by regions with sparse neuronal activity anda reduced background noise compared with that observedwithin the STN. Microinjections were performed in the samemicroelectrode track that would later be used to insert theDBS electrode (five patients). Microinjection in the onepatient undergoing a subthalamotomy was performed at alocation that was 3 mm anterior to the centre of a subsequentlesion (see Fig. 1A). Postoperative MRI was carried out infour patients to identify the placement of DBS electrodes orlesion in the STN and to confirm that microinjections wereperformed in the desired location in the STN (see Table 1).All patients underwent postoperative Unified Parkinson’sDisease Rating Score (UPDRS) assessment of the clinicalimprovement due to lesion or DBS (in the OFF drug state).

Simultaneous microinjection and microelectroderecording of neuronal activityIn two patients (Patients A and F), simultaneous micro-electrode recording of neuronal activity during microinjectionwas performed using a second independently drivenmicroelectrode that was inserted in parallel with the injectioncannula at a centre to centre distance of ~600 µm. Both themicroelectrode and the cannula occupied separate guide tubes(23 gauge stainless steel tubes) and the guide tubes werepositioned medial–lateral to one another. Microelectroderecordings were performed posterior–ventral to the injectionsite (i.e. microelectrode was advanced ahead of the injectionsite). Spectral analysis of single-neurone discharge activityhas previously been described in detail (Levy et al., 2000).Briefly, the event times of neuronal discharges were convertedto waveforms (with a final sampling rate of 1000 Hz)representing the discharge density over time (10 ms bins)using standard software (Spike2; Cambridge ElectronicDesign, Cambridge, UK). Subsequent Fourier analysis wasused to determine the oscillatory modulation of this discharge.Five hundred and twelve spectral estimates were madebetween 0 and 500 Hz thereby yielding a frequency resolutionof 0.98 Hz. Statistical significance of spectral peaks wasassessed by comparing spectra with a white noise signal withthe same mean power in the range 0–40 Hz (i.e. meanspectral ‘noise’). Spectral estimates were deemed significantif they were equal to or greater than the upper bound of a100√(1�α)% confidence interval about this white noisesignal. The upper bound was given by 2N√f�(ω)/χ2

2N, 1–α/2

where f is the power, ω is the frequency, 2N is the effectivedegrees of freedom and N is the number of sampling windows(Chatfield, 1996). A plot of frequency versus time wasconstructed by analysing data in consecutive non-overlapping

20-s windows. This plot displayed normalized spectra whereeach spectral estimate was divided by the spectral noise inthe respective window.

Assessment of clinical changes due tomicroinjectionsAll injections were performed with the patients in the OFFstate (12–14 h after last anti-parkinsonian medications). Inorder to rate the clinical effect of microinjections, patientsunderwent a partial UPDRS assessment of bradykinesia (item24), rigidity (item 22, arm/wrist) and tremor (items 20 and21, arm/wrist). Tremor, bradykinesia and dyskinesias wereassessed from camera tapes by one of the investigators(A.E.L.). Changes in rigidity were assessed by either oftwo investigators (A.E.L. or A.M.L.). When possible,bradykinesia was also assessed using quantitative tasksmeasuring movement time and amplitude such as wristpronation/supination or repetitive pointing with the indexfinger from the patient’s chest to a target placed ~50 cm infront of the patient (total trial lengths ~10 s). In one patient,rigidity was quantified using a commercially available device(Prochazka et al., 1997). This device allows the accuratedetermination of rigidity by quantifying the impedanceprovided by a limb to an applied force imposed by theexaminer.

ResultsLocalization of injectionsFigure 1 displays digitized parasagittal plates (Schaltenbrandand Wahren, 1977) of the STN (standardized to the patient’santerior–posterior commissural distance) to show the locationof the microelectrode trajectories and to provide informationabout the neuronal activity and the location of the individualmicroinjections. In all patients, microinjections wereperformed in regions of the STN populated by neuroneswith movement-evoked activity, tremor-related activity orneurones characteristic of those found in the motor portionof the STN (Hutchison et al., 1998). As shown in Table 1,STN DBS in five patients and the subthalamotomy inthe remaining patient (Patient A) resulted in a dramaticimprovement in contralateral bradykinesia, tremor andrigidity. Postoperative MRI was available in four patientsand confirmed that the interventions were performed inthe STN.

Lidocaine injectionsThe diffusion of lidocaine was examined in Patient A bysimultaneously recording neural activity close to the injectioncannula. These data are shown in Fig. 2. After 3.5 µl oflidocaine was injected over a 5 min period, there was adramatic decrease in neural activity recorded 0.6 mm fromthe injection site. When the recording electrode was moved

2108 R. Levy et al.

Fig. 1 Sagittal sections from the Schaltenbrand and Wahren stereotactic atlas at 12 mm from themidline displaying the microelectrode trajectories, neuronal activity and location of microinjection in thesix patients (definitions of the symbols used are at the bottom of the figure). DBS electrodes wereplaced in the same microelectrode trajectory as the microinjection in Patients B–F. The lesion in PatientA was placed 3 mm posterior to the location of the lidocaine microinjection. The diameter of theinjection cannula was 0.3 mm, but is drawn as 0.2 mm so as not to obscure data in the figure.Simultaneous recording of neuronal activity during microinjection was performed in Patients A and F(see Methods). STN � subthalamic nucleus; Thal � thalamus; SNr � substantia nigra pars reticulata;H2 � fields of Forel; rTh � reticular nucleus of the thalamus; Voa � nucleus ventralis oralis anterior;Vop � nucleus ventralis oralis posterior; Vim � ventrointermedius. The dashed line represents theanterior–posterior commissural line; RF � receptive field (a neurone that responded to passive or activelimb movement).

to another cell 0.78 mm away (relative to the tip of theinjection cannula), neural discharge at this site decreased~30 s later. However, neural activity 1.2 mm away was not

blocked at 5 min after the end of the injection period. Nochange in contralateral arm rigidity, tremor or bradykinesiawas observed at 13 min after the injection (data not shown).

STN microinjections in Parkinson’s disease patients 2109

Table 1 The clinical effects of microinjections and STN, DBS or subthalamotomy

Patient STN side Clear therapeutic Procedure* Contralateral Contralateral Contralateral Postoperative MRIbenefit of UPDRS UPDRS rigidity‡ UPDRS targetinginjection? tremor† bradykinesia§

A Left, 3.5 µl No Left STN 6/1 5/1 13/5 Centred in left STNlidocaine lesion

(subthalamotomy)(1 week)

B Right, 10 µl Yes Bilateral DBS 2/0 5/3 14/4 Not availablelidocaine (1 month)

C Right, 23 µl Yes Bilateral DBS 3/0 6/3 11/7 Effectivelidocaine (3 months) contacts in STN

D Left, 10 µl Yes Unilateral 7/0 Not available 14/10 Effective contactslidocaine DBS (3 weeks) in STN

E Left, 10 µl Yes Bilateral DBS 4/2 6/5 8/5 Effective contactsmuscimol (1 month) in STN

F Left, 5 µl Yes Bilateral DBS 4/1 4/0 5/1 Not availablemuscimol (3 weeks)

All patients were assessed in the OFF drug condition (following an overnight drug holiday). For Patients B–F, data are reported as theUPDRS score for DBS OFF/DBS ON (DBS at optimal parameters and contacts). For Patient A, data are reported as the UPDRS forpreoperative score/postoperative score. Scores are reported for the body side contralateral to the side of injection in the STN. *Stimulationused and postoperative time of assessment is indicated in parenthesis. †Tremor is calculated as the sum of UPDRS items 20 and 21(unilateral scores; maximum possible score of 12). ‡Rigidity is calculated from UPDRS item 22 (unilateral scores; maximum possible scoreof 8). §Bradykinesia is calculated as the sum of UPDRS items 23–26 (unilateral scores; maximum possible score of 16).

A marked anti-parkinsonian effect was observed in threepatients (Patients B, C and D). All these patients received�10 µl of lidocaine and the results are displayed in Fig. 3.In all three patients there was a decrease in rigidity and thisreached a minimum value ~10–20 min after the start ofinjection. Recovery of rigidity to baseline values was observedin two patients (Patients B and C). There were no ‘recovery’data available for Patient D because lidocaine was injectednear the end of the stereotaxic procedure. Lidocainemicroinjections into regions of the STN with tremor-relatedactivity reduced contralateral arm tremor in the two patientswith tremor during the procedure (Patients B and D).Accelerometer traces at the bottom of Fig. 3 show that thebaseline resting tremor of Patient D was intermittent withbrief (~10 s) periods of tremor. At 6 min, tremor amplitudewas decreased and the residual tremor was more intermittent.Patients B and D performed quantitative movement tasks.There was a significant decrease in the time required forPatient B to perform repetitive chest to button board pointingafter the injection (displayed in the upper right bar graph inFig. 3, see Methods). In Patient D, the amplitude of wristpronation/supination movements increased after lidocaineinjection (traces shown to the right of UPDRS ratings).

Following lidocaine injections, dyskinesias were observedin all three patients. Patient B developed low amplitudechoreoathetotic movements of the ipsilateral foot (indicatedby dashed arrow in Fig. 3). Patient C developed dystoniccontralateral wrist movements at 15 min after injection. Thesemovements interfered with the patient’s ability to performhand-gripping movements (filled squares in Fig. 1, middlepanel, UPDRS item 24). Patient D displayed dystonic

dyskinesias (i.e. a sustained posturing) of the contralateralfoot at 6 min following injection. At 9 min followinginjection, myoclonic jerking movements were also observed.Dyskinetic movements in the foot were also observed whenthe patient performed wrist pronation/supination movementsduring this period (11 min after initial injection).

Muscimol injectionsA marked anti-parkinsonian effect following muscimolinjections was observed in Patients E and F. Figure 4demonstrates that muscimol but not saline (both injected ata rate of 1 µl/min) injected into an area of the STNwith tremor-related activity in Patient E caused a dramaticreduction in contralateral resting tremor within 5 min afterthe start of injection. This was also observed using anaccelerometer placed on the index finger of the contralateralhand (bottom traces in Fig. 4). The patient was asked toperform mental arithmetic throughout the sampling period toenhance spontaneous resting tremor. Muscimol also improvedthe performance of wrist pronation/supination movements.

An injection of 5 µl of muscimol into a region with tremor-related activity also decreased resting tremor in Patient F.The changes in hand tremor and the discharge activity of aneurone that was located 1.3 mm from the injection site areshown in Fig. 5. A 5 Hz resting tremor of the hand wasmeasured with an accelerometer placed on the contralateralindex finger (Fig. 5A). There was no elbow tremor and pillrolling movements were not observed. There was no otherdetectable tremor on the side contralateral to the recording.The patient had a resting tremor of the ipsilateral hand/wrist

2110 R. Levy et al.

Fig. 2 Microelectrode recording of neuronal activity during themicroinjection of lidocaine in Patient A. Lidocaine was injected atthe volumes (µl, bold numbers in grey boxes) and times indicatedby the grey boxes to the left side of vertical timeline. The relativedistance from the tip of the injection cannula to the tip ofmicroelectrode is indicated by the numbers on the left hand sideof the figure. Neuronal recordings were sampled at the timeindicated by position of traces with respect to timeline.

and ipsilateral foot. Changes in rigidity and bradykinesiawere not assessed in this patient because of the need tomaintain a stable microelectrode recording from the singleneurone. Microstimulation along this microelectrodetrajectory produced tremor arrest in the contralateral hand(1 mm above to 3 mm below injection site in Fig. 1). Theleft panel of Fig. 5B demonstrates that before the injection ofmuscimol, the neurone displayed two peaks in the frequency

spectra, one peak at 4 Hz and another at 20 Hz. The lowerfrequency oscillatory component of the recorded neuronewas not coherent with the contralateral wrist flexor EMG orthe accelerometer signal at any time (not shown). Followingthe injection of muscimol, the lower frequency oscillatorycomponent decreased concurrent with a reduction incontralateral hand tremor (Fig. 5B, right panel). No changewas observed for ipsilateral tremor. Interestingly, there wasno suppression in the firing rate of the neurone (see Fig. 5A)or the 20 Hz oscillatory activity. Time-frequency analysisrevealed that the reduction of 4 Hz oscillatory neuronaldischarge (Fig. 5B, right panel) was due to a decrease in alloscillatory activity �10 Hz rather than an increase in themoment-to-moment variability of a peak within this frequencyband. Contralateral hand tremor reappeared ~5 min after theend of the injection period and, at that time, the recordedneurone started to display a lower frequency oscillatorycomponent in addition to the 20 Hz high-frequency oscillatorycomponent.

DiscussionInactivation of the STN improves parkinsonismThis is the first report of microinjections of muscimol andlidocaine in the STN of humans. This study demonstratesthat local inactivation of the STN using pharmacologicalblocking agents results in a transient improvement in akinesia,rigidity and limb tremor in patients with Parkinson’s disease.These results support the current model of basal gangliapathophysiology, which predicts that a reduction of excessiveactivity in the STN of patients with Parkinson’s diseaseshould produce a therapeutic benefit (DeLong, 1990). Theseresults in humans are consistent with previous observationsmade in the non-human primate MPTP model of parkinsonism(Bergman et al., 1990; Aziz et al., 1991, 1992; Guridi et al.,1994; Wichmann et al., 1994).

Suppression of cell bodies and axons versussuppression of only cell bodiesAlthough subthalamotomy and STN DBS in patients withParkinson’s disease improves all parkinsonian cardinal motorsigns (Limousin et al., 1995; Gill and Heywood, 1997; Obesoet al., 1997; Alvarez et al., 2001), it is unknown if theseeffects are due solely to inactivation of STN neurones ratherthan an action on fibres passing through or near the affectedregion. The anti-parkinsonian effect of muscimol in thisstudy confirms that suppression of STN sensorimotor-relatedneuronal activity, as opposed to possible alterations in nearbypallidofugal fibres, results in a therapeutic benefit.

Since lidocaine acts on fibres passing through the blockedvolume of cell bodies, it can indirectly affect other STNregions by preventing transmission through both inhibitoryand excitatory afferent inputs and excitatory efferentprojections. Although inhibition of GABAergic afferent fibres

STN microinjections in Parkinson’s disease patients 2111

Fig. 3 The effects of lidocaine injections in the STN of three patients. The three plots show the effectsof lidocaine injection on the UPDRS measures of tremor, rigidity and hand grips and the rigidity meterreadings (see box at bottom right for distinction of graph symbols, values for the body side contralateralto the injected STN). The stippled bars show time period over which lidocaine was injected, the volumeof lidocaine in µl is given by the numbers in the bars. Lidocaine resulted in a decrease in rigidity in allthree patients and the rigidity in two of these patients (Patients B and C) was observed to recover tobaseline values. Lidocaine reduced limb tremor in Patients B and D (Patient C did not have limb tremorduring the procedure). Traces at the bottom of the figure are accelerometer traces of 90 s of restingtremor of Patient D (traces were sampled starting at the times indicated to the left). The bar graph atthe top right shows that lidocaine injections in Patient B resulted in improved performance on amovement time task (see Methods, *P � 0.05, ANOVA on ranks). The traces at the bottom right showthe improvement in repetitive wrist pronation/supination movements (WPS) in Patient D followinglidocaine injections. The time of appearance of dyskinetic movements is indicated by downward dashedarrows. Following lidocaine injections, all three patients developed dystonic or dyskinetic movements(see Results).

2112 R. Levy et al.

Fig. 4 The effect of muscimol injection in the STN on clinical symptoms in Patient E. The plot at thetop left shows the effect of muscimol on bilateral resting tremor and rigidity (UPDRS scores, see box atbottom right for distinction of graph symbols). The stippled bar (muscimol) and the white bar (saline)show time period over which each substance was injected and the volume of substance in µl in thebars. The plot shows that contralateral but not ipsilateral UPDRS resting tremor was reduced by amuscimol injection into an area of the STN with tremor-related activity (see Fig. 1). Muscimol alsoimproved the performance of wrist pronation/supination (WPS) movements (traces at top right). Thetraces at the bottom of the figure demonstrate that muscimol but not saline reduced spontaneous restingtremor recorded with an accelerometer (see Results).

from the globus pallidus externus (GPe) would be predictedto enhance STN activity (Rouzaire-Dubois et al., 1980), ithas been suggested that the increased activity of the STN inParkinson’s disease is due primarily to excitatory drive(Hassani et al., 1996; Levy et al., 1997). As well as inputfrom GPe, the STN receives massive glutamatergic inputfrom the cerebral cortex (Monakow et al., 1978; Rouzaire-Dubois and Scarnati, 1985; Canteras et al., 1990; Nambuet al., 1996), forms reciprocal projections with thepedunculopontine nucleus (Hammond et al., 1983; Lavoieand Parent, 1994) and receives input from the centromedian–parafascicular complex of the thalamus (Sadikot et al., 1992;Mouroux et al., 1995; Orieux et al., 2000). Inhibition bylidocaine of these excitatory afferent projections to otherregions of the STN, in addition to the inhibition of justSTN neurones, would also be predicted to have an anti-parkinsonian effect. Equivalently, inhibition of efferent axonalprojections from other STN neurones (whose somata are notwithin the blocked area) by lidocaine would also act todecrease glutamatergic excitatory outflow from the STN.Therefore, it is possible that the functional volume of theSTN that is blocked using lidocaine is larger than the volumeof the affected cell bodies. Since muscimol specificallyinhibits cell bodies, the anti-parkinsonian effects of muscimolobserved in this study are probably due to the focal inhibitionof excitatory output from the blocked region only.

DyskinesiasThe involvement of the STN in the pathogenesis ofdyskinesias is supported by many studies (Crossman, 1990).Dyskinesias and dystonic postures are observed in normalmonkeys following excitotoxic cell-specific lesions (Hamadaand DeLong, 1992) or electrical stimulation of the STN(Beurrier et al., 1997). Dyskinesias can also be induced byinjection of the GABA antagonist bicuculline in the GPe(Matsumura et al., 1995) and disinhibition of the GPe shouldlead to a decrease in STN activity. In MPTP-treated monkeys,inactivation of the STN results in transient dyskinesias inaddition to improving parkinsonism (Bergman et al., 1990;Aziz et al., 1991, 1992; Guridi et al., 1994; Wichmannet al., 1994). In patients with Parkinson’s disease, choreic ordystonic dyskinesias are sometimes observed duringpenetration of the STN target by microelectrodes or DBSelectrodes or during lesion making (Benabid et al., 2000;Alvarez et al., 2001).

Limousin and colleagues demonstrated that STN DBS athigher voltages than those required to control parkinsonismcould also induce dyskinesias in patients with Parkinson’sdisease (Limousin et al., 1996). It was suggested that becausethe anti-parkinsonian effect and dyskinetic effect occurredwith different stimulation parameters, the mechanismsresponsible for these two effects might be distinct. In thepresent study, �10 µl of lidocaine reduced parkinsonian

STN microinjections in Parkinson’s disease patients 2113

Fig. 5 Effects of muscimol injection in the STN of Patient F. (A) The top trace shows the accelerometer (Acc) recording of the hand/wrist tremor (accelerometer was placed on index finger). The histogram shows the firing rate (20 s bins) of an STN neurone recordedduring the injection and located 1.3 mm from the cannula tip. The grey bar represents the time period over which the 5 µl of muscimolwas injected. (B) Power spectra of oscillatory neuronal discharge recorded before (left panel, 0–3 min) and after (right panel, 9–12 min)muscimol injection (0.24 Hz resolution). The inset in each plot shows the frequency spectra of the accelerometer over each time period(0.39 Hz resolution). (C) Time-frequency analysis of oscillatory neuronal discharge (0.98 Hz resolution � 20 s bins). The colour legendindicates those spectral estimates with signal-to-noise ratios (green to light green) that were greater than a 99.8% confidence intervalabout the mean spectral noise. These plots demonstrate that the lower frequency oscillatory activity of the recorded neurone, but notfiring rate or 20 Hz oscillatory activity, was reduced concurrent with a reduction in hand tremor.

2114 R. Levy et al.

symptoms and produced dyskinesias in three patients, whilein the remaining patient (Patient A), no motor benefit wasobtained and no dyskinesias were observed following aninjection of 3.5 µl of lidocaine. These results suggest that,in parkinsonian patients, the dyskinetic effect requires agreater degree of STN inactivation than does the anti-parkinsonian effect. This would explain why injections ofmuscimol did not produce dyskinesias in this study but didproduce an anti-parkinsonian effect if the volume of the STNinactivated by muscimol produced only a threshold anti-parkinsonian effect, but was not large enough to producedyskinesias. It is also of interest that both choreic anddystonic dyskinesias were observed in the present study.These types of dyskinesias were typical of those seen duringlevodopa-induced dyskinesias (Fahn, 2000) and support thefinding that the neuronal activity of the STN in patients withParkinson’s disease is reduced during apomorphine-induceddyskinesias (apomorphine is a non-selective D1- andD2-dopamine receptor agonist) (Levy et al., 2001).

Location and spread of microinjectionsThe location of microinjections into the STN in this studywas determined using microelectrode techniques andpostoperative MRI verification of DBS electrode or lesionposition relative to the STN. In addition, DBS in thesame microelectrode tracks as the injections resulted inan improvement in tremor, rigidity and bradykinesia. Thesimultaneous recording of neuronal activity in two patientswas also used to confirm that injections were performed inthe STN. This technique is the most direct way to assess theeffectiveness and spread of inactivating agents (Malpeli,1999), and was especially important with regard to PatientA because it demonstrated that the lidocaine injection waswell targeted and that the injection effectively inactivatedneuronal tissue although no clinical effect was observed.

By recording neuronal activity at multiple distances fromthe injection site, we were also able to assess the spread oflidocaine and the size of the effective block. It took ~5 minfor 3.5 µl of lidocaine to block neuronal activity within~1 mm of the injection site in the posterior–ventral direction(i.e. below the cannula, see Fig. 1). Our results are comparablewith those of Martin and Ghez (Martin and Ghez, 1999).They demonstrated, using autoradiographic monitoring ofglucose uptake/metabolism, that following an injection of1 µl of lidocaine a reduction in glucose uptake within ~1 mmof the injection centre was attributable to drug spread. Otherobservations of diffusion distances in rat thalamic and spinaltissue have demonstrated that a 1 µl injection of lidocainewould result in a block of neuronal activity in a sphericalregion of radius 0.8 mm at 10 min after injection (Myers,1966; Sandkuhler et al., 1987). Although we injected 3.5 µland these studies injected 1 µl of lidocaine, there are manyfactors that could account for variability in diffusion distances.For example, it is assumed that the distribution of lidocaineis spherical in shape in nuclear and cortical regions (Myers,

1966; Martin and Ghez, 1999), but there is significantanisotropy in the direction and extent of diffusion fromthe injection site in regions containing fibres of passage(Sandkuhler et al., 1987). Some of the injected lidocainemight have also preferentially diffused up along the cannulashaft and therefore reduced the diffusion distances observedwith our simultaneous microelectrode recording set-up. Hupeand colleagues demonstrated that with pressure injections,inactivation is more efficient above than below the pipettetip and that the volume of inactivation has an ellipsoidalform centred above the tip of the pipette (Hupe et al., 1999).Other factors contributing to the observed differences indiffusion distances include: the volume of the extracellularspace, the diffusibility of the drug through the extracellularspace, the homogeneity of the medium of diffusion (whitematter versus grey matter), the vascularization of the tissuesurrounding the injection cannula (which would affect thewashout of the drug), how fast the drug is degraded and therate of drug delivery (Sandkuhler et al., 1987; Malpeli, 1999).

One caveat of this study is the inability to completely ruleout that the observed effects were not due to diffusion oflidocaine or muscimol into neighbouring structures(especially when simultaneous recording was not performed),most notably the substantia nigra pars reticulata. The effectsof these agents on this nearby structure might have a similareffect to STN inactivation since increased neuronal activityand abnormal patterning of the substantia nigra pars reticulatais observed in the MPTP monkey model of Parkinson’sdisease (Wichmann et al., 1999).

Rates of injection and volume effectsAlthough the injection rates used in this study (1–2.5 µl/min)were greater than those used in animals (0.1–2 µl/min)(Myers, 1966; Demer and Robinson, 1982; Duncan et al.,1993; Wichmann et al., 1994; Burbaud et al., 1998), thereare several factors that indicate that volume effects didnot produce the observed clinical effects. (i) Simultaneousrecording of neuronal activity was quite stable during theperiod of injection and indicates that tissue was not beingdeformed at distances �0.6 mm from the injection site whensubstances were injected at a rate of ~1 µl/min. (ii) It isunlikely that the effects observed were due to a volumeeffect since there was a delay in the anti-parkinsonian effectproduced by these injections; e.g. in Patient C (23 µl injectedat 2.5 µl/min), a marked decrease in rigidity did not occuruntil 20 min following injection. In contrast, the subsequentinsertion of the DBS electrode in this patient resulted in animmediate improvement in rigidity (not shown) [NB if themaximum length of the DBS electrode (1.27 mm in diameter)in the STN is ~7 mm (see Fig. 1), the volume of the STNtissue displaced by the DBS electrode is ~9 µl]. Benabid andco-workers have also reported that a significant decrease inakinesia and rigidity, along with the emergence of ballisticor choreoballistic movements, occurs at the time of insertionof the DBS electrode in patients with Parkinson’s disease

STN microinjections in Parkinson’s disease patients 2115

(Benabid et al., 2000). However, Demer and Robinsondemonstrated that injections of �12 µl and/or at rates�1 µl/min produced irreversible damage in a region within~0.7 mm from the injection site, which was marked by gliosisand fibre loss (Demer and Robinson, 1982). Therefore, wecannot rule out that tissue damage did not occur due tohigher rates of injection (especially in Patient C). Yet, it isunlikely that the observed clinical effects were due topermanent tissue damage because there was a return tobaseline parkinsonism in all patients in whom recovery datawere available.

Time courseThe time course of the clinical effects of the lidocaine andmuscimol injections, beginning 5–10 min following injectionand lasting 30–50 min, closely matched those seen in animals(Sandkuhler et al., 1987; Wichmann et al., 1994; Martinand Ghez, 1999). In contrast, recovery to baseline tremorfollowing a 5 µl muscimol injection (Patient F) was observed~5 min after the end of the injection. This time course issimilar to the time course of tremor suppression observedfollowing injections of a similar volume of muscimol in thethalamus of patients with essential tremor (Pahapill et al.,1999). These results are consistent with studies using GABAwhich demonstrate that the duration of neuronal inactivationis proportional to the volume of substance injected (Hupeet al., 1999). These data suggest that the potential clinicalapplication of microinjections to aid in the determination ofthe optimal target for lesions or DBS is limited. A significantadvantage of the use of stimulation techniques over the useof microinjections as described in the present work is thatintraoperative stimulation decreases tremor or rigidity witha very short latency (�1 min) and is highly reproducible(Rodriguez et al., 1998).

Inhibition of tremor-related neuronal activitysuppresses limb tremorFollowing MPTP-treatment in monkeys, there is a prominentincrease in the number of STN neurones that displayoscillatory activity (Bergman et al., 1994), and limb tremorreduction has been demonstrated following muscimolinjections in the STN (Wichmann et al., 1994). In the presentstudy, injections of muscimol into regions of the STN withtremor-related activity caused a reduction in limb tremor.Our results support the view that oscillatory activity in theSTN is important in the mediation of parkinsonian limbtremor (Bergman et al., 1994).

It is interesting that the latency of the effect of muscimolor lidocaine on limb tremor was short. Five minutes afterthe start of a 5 µl lidocaine injection, we observed a reductionof limb tremor in Patients B and D, and in Patients E and Fafter 5 µl of muscimol. Although we did not wait for a longperiod after the saline injection in Patient E, it was observed

that saline did not affect tremor. The effect of muscimoloccurred immediately after 5 µl of muscimol was injected,while there was no immediate effect of saline (both wereinjected at 1 µl/min). Since tremor suppression occurred withsmall volumes of inactivating agents and with a shortlatency, our results suggest that tremor suppression can beaccomplished by silencing the activity of a small region ofSTN containing tremor cells. Evidence from simultaneousmicroelectrode recording techniques has indicated thatpotential tremor-generating circuits in the basal ganglia canoccupy a relatively small volume (�1 mm3) (Hurtado et al.,1999; Levy et al., 2000). Our results are also consistent withthose of Rodriguez and colleagues who demonstrated thatmicrostimulation in STN regions containing tremor cellsreduces tremor with a very short latency (Rodriguez et al.,1998). It has been suggested that microstimulation affects asmall volume of tissue at currents �100 µA, the maximumgenerally possible with microelectrodes (Dostrovsky et al.,2000).

Changes in tremor-related oscillations beyondthe blocked region of tremor cellsThere is now much evidence from single-unit microelectroderecordings indicating the involvement of the STN in thepathogenesis of parkinsonian limb tremor (Bergman et al.,1990, 1994; Hutchison et al., 1998; Krack et al., 1998b;Rodriguez et al., 1998; Magarinos-Ascone et al., 2000).However, increasing our understanding of thepathophysiology of tremor and expanding the current modelof Parkinson’s disease has required the use of novelexperimental techniques. To this end, simultaneous recordingof tremor-related activity at multiple sites has provedinvaluable (Nini et al., 1995; Bergman et al., 1998a, b;Hurtado et al., 1999; Levy et al., 1999, 2000; Raz et al.,2000). In the present study, simultaneous recording ofneuronal activity during microinjection allowed us to directlyassess the time course of the effect of a pharmacologicalblock in an area not directly deactivated by the block. Wedemonstrated that the 4 Hz discharge oscillation in a neuronelocated outside a blocked region of tremor cells was reducedconcurrent with a reduction in limb tremor. However, thefiring rate and 20 Hz oscillatory activity remained unchanged.This result suggests that a block of neuronal activity canmodify the neuronal discharge of cells located beyond theblocked area. There are several possible ways that this resultmight be interpreted.

First, if the recorded neurone received proprioceptive inputfrom the tremulous contralateral hand/wrist, reduction oflimb tremor could account for a reduction in the lowerfrequency oscillatory activity of the recorded neurone (Kracket al., 1998b). However, at no time was the oscillatoryactivity of the neurone coherent with wrist flexor EMG orthe accelerometer signal, which would be expected if itsfiring was dependent on sensory input. Secondly, if the

2116 R. Levy et al.

recorded neurone received 4 Hz oscillatory input directlyfrom cells at the injection site, inactivation of these cellswould also lead to a reduction in the lower frequencycomponent of the recorded neurone (Martin and Ghez, 1999).Yet, although intrinsic axon collaterals in the STN arefrequent in the rat (Kita et al., 1983), they are rare in primates(Yelnik and Percheron, 1979; Sato et al., 2000). Furthermore,coincident discharge between pairs of STN neurones inpatients with Parkinson’s disease is not observed, suggestingthat STN tremor cells do not directly drive the activity ofother STN tremor cells (Levy et al., 2000). Thirdly, muscimolmay have affected the activity of the recorded neurone byacting on a dendrite extending within the blocked region.For example, it has been shown that the dendritic fields ofneurones located centrally in the STN can extend over nearlytwo-thirds of the structure (Sato et al., 2000). However, thiswould probably result also in a decrease in the overallspontaneous discharge of the neurone, and this was notobserved. Fourthly, both parkinsonian limb tremor (Scholzand Bacher, 1995) and tremor-related activity in the basalganglia (Bergman et al., 1998b; Hurtado et al., 1999; Levyet al., 2000; Raz et al., 2000) have been shown to exhibitfluctuations in rhythmic activity. However, it is unlikely thatboth the decrease in 4 Hz oscillations of the recorded neuroneand a decrease in 5 Hz limb tremor are merely due tospontaneous variability, because these oscillations decreasedand reappeared with a similar time course.

Lastly, suppression of limb tremor could arise as a resultof, in addition to direct inactivation of tremor-relateddischarge, the ‘desynchronizing’ effect of a small block ofthe STN upon neighbouring subcircuits in the corticobasalganglia–thalamic loop (Bergman et al., 1998a; Deuschl et al.,2000). This concept is based on the observation that increasedtremor-related oscillatory synchronization in the basal gangliaunderlies the development of limb tremor in MPTP-treatedmonkeys (Bergman et al., 1998a; Raz et al., 2000). It hasalso been hypothesized that the STN might play a rolein synchronizing oscillatory activity in the GPi, becauseoscillations in the tremor frequency range within the GPipersist in spite of a reduction in limb tremor following lesionsof the STN (Wichmann et al., 1994). The reduction ofoscillatory activity in the tremor-frequency range of a neuronelocated beyond a deactivated region, as demonstrated in thepresent study, supports these hypotheses. It is interesting tonote that high-frequency oscillatory activity was notsuppressed during the effect of the muscimol block, and thisfurther supports the notion that the underlying mechanismsof the tremor and high-frequency oscillations in the STN aredifferent (Levy et al., 2000).

SummaryThis study demonstrates that microinjections of pharmaco-logical blocking agents in the STN of patients withParkinson’s disease results in a transient anti-parkinsonianeffect, similar to results from animal models of Parkinson’s

disease. Consistent with predictions of the current model ofParkinson’s disease (DeLong, 1990), inactivation of theSTN also produces dyskinesias. In addition to suppressingoscillatory STN output, inactivation of neurones with tremor-related activity possibly results in a reduction of limb tremordue to the desynchronizing effect of the focal block onpathological oscillations in the basal ganglia (Bergman et al.,1998a; Deuschl et al., 2000).

AcknowledgementsWe wish to thank the patients for their participation. Wegratefully acknowledge the assistance of Eppie Sime RNfrom the Movement Disorders Center at the Toronto WesternHospital. Funding was provided by the US National Instituteof Health, Canadian Institute of Health Research and theParkinson’s Foundation of Canada. A.M.L. is a MedicalResearch Council of Canada clinician scientist.

ReferencesAlvarez L, Macias R, Guridi J, Lopez G, Alvarez E, Maragoto C,et al. Dorsal subthalamotomy for Parkinson’s disease. Mov Disord2001; 16: 72–8.

Aziz TZ, Peggs D, Sambrook MA, Crossman AR. Lesion of thesubthalamic nucleus for the alleviation of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-induced parkinsonism in the primate.Mov Disord 1991; 6: 288–92.

Aziz TZ, Peggs D, Agarwal E, Sambrook MA, Crossman AR.Subthalamic nucleotomy alleviates parkinsonism in the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-exposed primate. Br JNeurosurg 1992; 6: 575–82.

Barlas O, Hanagasi HA, Imer M, Sahin HA, Sencer S, EmreM. Do unilateral ablative lesions of the subthalamic nucleus inparkinsonian patients lead to hemiballism? Mov Disord 2001; 16:306–10.

Benabid AL, Benazzouz A, Limousin P, Koudsie A, Krack P, PiallatB, et al. Dyskinesias and the subthalamic nucleus. [Review]. AnnNeurol 2000; 47 (4 Suppl 1): S189–92.

Bergman H, Wichmann T, DeLong MR. Reversal of experimentalparkinsonism by lesions of the subthalamic nucleus. Science 1990;249: 1436–8.

Bergman H, Wichmann T, Karmon B, DeLong MR. The primatesubthalamic nucleus. II. Neuronal activity in the MPTP model ofparkinsonism. J Neurophysiol 1994; 72: 507–20.

Bergman H, Feingold A, Nini A, Raz A, Slovin H, Abeles M, et al.Physiological aspects of information processing in the basal gangliaof normal and parkinsonian primates. [Review]. Trends Neurosci1998a; 21: 32–8.

Bergman H, Raz A, Feingold A, Nini A, Nelken I, Hansel D, et al.Physiology of MPTP tremor. Mov Disord 1998b; 13 Suppl 3: 29–34.

Beurrier C, Bezard E, Bioulac B, Gross C. Subthalamic stimulationelicits hemiballismus in normal monkey. Neuroreport 1997; 8:1625–9.

STN microinjections in Parkinson’s disease patients 2117

Brown P, Oliviero A, Mazzone P, Insola A, Tonali P, Di LazzaroV. Dopamine dependency of oscillations between subthalamicnucleus and pallidum in Parkinson’s disease. J Neurosci 2001; 21:1033–8.

Burbaud P, Bonnet B, Guehl D, Lagueny A, Bioulac B. Movementdisorders induced by gamma-aminobutyric agonist and antagonistinjections into the internal globus pallidus and substantia nigra parsreticulata of the monkey. Brain Res 1998; 780: 102–7.

Canteras NS, Shammah-Lagnado SJ, Silva BA, Ricardo JA. Afferentconnections of the subthalamic nucleus: a combined retrograde andanterograde horseradish peroxidase study in the rat. Brain Res 1990;513: 43–59.

Chatfield C. The analysis of time series: an introduction. 5th ed.London: Chapman & Hall; 1996.

Crossman AR. A hypothesis on the pathophysiological mechanismsthat underlie levodopa- or dopamine agonist-induced dyskinesia inParkinson’s disease: implications for future strategies in treatment.[Review]. Mov Disord 1990; 5: 100–8.

DeLong MR. Primate models of movement disorders of basalganglia origin. [Review]. Trends Neurosci 1990; 13: 281–5.

Demer JL, Robinson DA. Effects of reversible lesions andstimulation of olivocerebellar system on vestibuloocular reflexplasticity. J Neurophysiol 1982; 47: 1084–107.

Deuschl G, Raethjen J, Baron R, Lindemann M, Wilms H, KrackP. The pathophysiology of parkinsonian tremor: a review. [Review].J Neurol 2000; 247 Suppl 5: V33–48.

Dostrovsky JO, Levy R, Wu JP, Hutchison WD, Tasker RR, LozanoAM. Microstimulation-induced inhibition of neuronal firing inhuman globus pallidus. J Neurophysiol 2000; 84: 570–4.

Duncan GH, Bushnell MC, Oliveras JL, Bastrash N, Tremblay N.Thalamic VPM nucleus in the behaving monkey. III. Effects ofreversible inactivation by lidocaine on thermal and mechanicaldiscrimination. J Neurophysiol 1993; 70: 2086–96.

Fahn S. The spectrum of levodopa-induced dyskinesias. [Review].Ann Neurol 2000; 47 (4 Suppl 1): S2–9.

Gill SS, Heywood P. Bilateral dorsolateral subthalamotomy foradvanced Parkinson’s disease. Lancet 1997; 350: 1224.

Guridi J, Herrero MT, Luquin R, Guillen J, Obeso JA.Subthalamotomy improves MPTP-induced parkinsonism inmonkeys. Stereotact Funct Neurosurg 1994; 62: 98–102.

Guridi J, Herrero MT, Luquin MR, Guillen J, Ruberg M, LagunaJ, et al. Subthalamotomy in parkinsonian monkeys. Behaviouraland biochemical analysis. Brain 1996; 119: 1717–27.

Hamada I, DeLong MR. Excitotoxic acid lesions of the primatesubthalamic nucleus result in transient dyskinesias of thecontralateral limbs. J Neurophysiol 1992; 68: 1850–8.

Hammond C, Rouzaire-Dubois B, Feger J, Jackson A, CrossmanAR. Anatomical and electrophysiological studies on the reciprocalprojections between the subthalamic nucleus and nucleus tegmentipedunculopontinus in the rat. Neuroscience 1983; 9: 41–52.

Hassani OK, Mouroux M, Feger J. Increased subthalamic neuronalactivity after nigral dopaminergic lesion independent of disinhibitionvia the globus pallidus. Neuroscience 1996; 72: 105–15.

Hupe JM, Chouvet G, Bullier J. Spatial and temporal parametersof cortical inactivation by GABA. J Neurosci Methods 1999; 86:129–43.

Hurtado JM, Gray CM, Tamas LB, Sigvardt KA. Dynamics oftremor-related oscillations in the human globus pallidus: a singlecase study. Proc Natl Acad Sci USA 1999; 96: 1674–9.

Hutchison WD, Allan RJ, Opitz H, Levy R, Dostrovsky JO, LangAE, et al. Neurophysiological identification of the subthalamicnucleus in surgery for Parkinson’s disease. Ann Neurol 1998; 44:622–8.

Kita H, Chang HT, Kitai ST. The morphology of intracellularlylabeled rat subthalamic neurons: a light microscopic analysis.J Comp Neurol 1983; 215: 245–57.

Krack P, Pollak P, Limousin P, Hoffmann D, Xie J, Benazzouz A,et al. Subthalamic nucleus or internal pallidal stimulation in youngonset Parkinson’s disease. Brain 1998a; 121: 451–7.

Krack P, Benazzouz A, Pollak P, Limousin P, Piallat B, HoffmannD, et al. Treatment of tremor in Parkinson’s disease by subthalamicnucleus stimulation. Mov Disord 1998b; 13: 907–14.

Kumar R, Lozano AM, Montgomery E, Lang AE. Pallidotomy anddeep brain stimulation of the pallidum and subthalamic nucleus inadvanced Parkinson’s disease. Mov Disord 1998; 13 Suppl 1: 73–82.

Lavoie B, Parent A. Pedunculopontine nucleus in the squirrelmonkey: projections to the basal ganglia as revealed by anterogradetract-tracing methods. J Comp Neurol 1994; 344: 210–31.

Levy R, Hazrati LN, Herrero MT, Vila M, Hassani OK, MourouxM, et al. Re-evaluation of the functional anatomy of the basalganglia in normal and parkinsonian states. [Review]. Neuroscience1997; 76: 335–43.

Levy R, Davis KD, Hutchison WD, Pahapill PA, Lozano AM,Tasker RR, et al. Simultaneously recorded neuron pairs in the motorthalamus of patients with Parkinson’s disease and essential tremor[abstract]. Soc Neurosci Abstr 1999; 25: 1408.

Levy R, Hutchison WD, Lozano AM, Dostrovsky JO. High-frequency synchronization of neuronal activity in the subthalamicnucleus of parkinsonian patients with limb tremor. J Neurosci 2000;20: 7766–75.

Levy R, Dostrovsky JO, Lang AE, Sime E, Hutchison WD, LozanoAM. The effects of apomorphine on subthalamic nucleus and globuspallidus internus neurons in patients with Parkinson’s disease.J Neurophysiol 2001; 86: 249–60.

Limousin P, Pollak P, Benazzouz A, Hoffmann D, Le Bas JF,Broussolle E, et al. Effect of parkinsonian signs and symptoms ofbilateral subthalamic nucleus stimulation. Lancet 1995; 345: 91–5.

Limousin P, Pollak P, Hoffmann D, Benazzouz A, Perret JE, BenabidAL. Abnormal involuntary movements induced by subthalamicnucleus stimulation in parkinsonian patients. Mov Disord 1996; 11:231–5.

Magarinos-Ascone CM, Figueiras-Mendez R, Riva-Meana C,Cordoba-Fernandez A. Subthalamic neuron activity related to tremorand movement in Parkinson’s disease. Eur J Neurosci 2000; 12:2597–607.

2118 R. Levy et al.

Magill PJ, Bolam JP, Bevan MD. Relationship of activity in thesubthalamic nucleus-globus pallidus network to corticalelectroencephalogram. J Neurosci 2000; 20: 820–33.

Malpeli JG. Reversible inactivation of subcortical sites by druginjection. [Review]. J Neurosci Methods 1999; 86: 119–28.

Marsden JF, Limousin-Dowsey P, Ashby P, Pollak P, Brown P.Subthalamic nucleus, sensorimotor cortex and muscleinterrelationships in Parkinson’s disease. Brain 2001; 124: 378–88.

Martin JH, Ghez C. Pharmacological inactivation in the analysis ofthe central control of movement. [Review]. J Neurosci Methods1999; 86: 145–59.

Matsumura M, Tremblay L, Richard H, Filion M. Activity ofpallidal neurons in the monkey during dyskinesia induced byinjection of bicuculline in the external pallidum. Neuroscience 1995;65: 59–70.

Miller WC, DeLong MR. Altered tonic activity of neurons in theglobus pallidus and subthalamic nucleus in the primate MPTPmodel of parkinsonism. In: Carpenter MB, Jayaraman A, editors.The basal ganglia II. Structure and function: current concepts. NewYork: Plenum Press; 1987. p. 415–27.

Monakow KH, Akert K, Kunzle H. Projections of the precentralmotor cortex and other cortical areas of the frontal lobe to thesubthalamic nucleus in the monkey. Exp Brain Res 1978; 33:395–403.

Mouroux M, Hassani OK, Feger J. Electrophysiological study ofthe excitatory parafascicular projection to the subthalamic nucleusand evidence for ipsi- and contralateral controls. Neuroscience 1995;67: 399–407.

Muthane UB, Varma RM, Sundarajan P, Hegde T, JayakumarPN, Shankar SK. Percutaneous trans-foramen ovale approach tosubthalamic nucleus (Varma’s technique): a shorter and economicaltechnique in the surgical management of Parkinson’s disease[abstract]. Mov Disord 1998; 13 Suppl 2: 71.

Myers RD. Injection of solutions into cerebral tissue: relationbetween volume and diffusion. Physiol Behav 1966; 1: 171–4.

Nambu A, Takada M, Inase M, Tokuno H. Dual somatotopicalrepresentations in the primate subthalamic nucleus: evidence forordered but reversed body-map transformations from the primarymotor cortex and the supplementary motor area. J Neurosci 1996;16: 2671–83.

Nini A, Feingold A, Slovin H, Bergman H. Neurons in the globuspallidus do not show correlated activity in the normal monkey, butphase-locked oscillations appear in the MPTP model ofparkinsonism. J Neurophysiol 1995; 74: 1800–5.

Obeso JA, Alvarez LM, Macias RJ, Guridi J, Teijeiro J, Juncos JL,et al. Lesion of the subthalamic nucleus (STN) in Parkinson’sdisease (PD) [abstract]. Neurology 1997; 48 (3 Suppl 2): A138.

Orieux G, Francois C, Feger J, Yelnik J, Vila M, Ruberg M, et al.Metabolic activity of excitatory parafascicular and pedunculopontineinputs to the subthalamic nucleus in a rat model of Parkinson’sdisease. Neuroscience 2000; 97: 79–88.

Pahapill PA, Levy R, Dostrovsky JO, Davis KD, Rezai AR, TaskerRR, et al. Tremor arrest with thalamic microinjections of muscimolin patients with essential tremor. Ann Neurol 1999; 46: 249–52.

Plenz D, Kital ST. A basal ganglia pacemaker formed by thesubthalamic nucleus and external globus pallidus. Nature 1999;400: 677–82.

Prochazka A, Bennett DJ, Stephens MJ, Patrick SK, Sears-Duru R,Roberts T, et al. Measurement of rigidity in Parkinson’s disease.Mov Disord 1997; 12: 24–32.

Raz A, Vaadia E, Bergman H. Firing patterns and correlations ofspontaneous discharge of pallidal neurons in the normal andthe tremulous 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine vervetmodel of parkinsonism. J Neurosci 2000; 20: 8559–71.

Rodriguez MC, Guridi OJ, Alvarez L, Mewes K, Macias R, VitekJ, et al. The subthalamic nucleus and tremor in Parkinson’s disease.Mov Disord 1998; 13 Suppl 3: 111–18.

Rouzaire-Dubois B, Scarnati E. Bilateral corticosubthalamic nucleusprojections: an electrophysiological study in rats with chroniccerebral lesions. Neuroscience 1985; 15: 69–79.

Rouzaire-Dubois B, Hammond C, Hamon B, Feger J.Pharmacological blockade of the globus pallidus-induced inhibitoryresponse of subthalamic cells in the rat. Brain Res 1980; 200: 321–9.

Ryan LJ, Sanders DJ, Clark KB. Auto- and cross-correlation analysisof subthalamic nucleus neuronal activity in neostriatal- and globuspallidal-lesioned rats. Brain Res 1992; 583: 253–61.

Sadikot AF, Parent A, Francois C. Efferent connections of thecentromedian and parafascicular thalamic nuclei in the squirrelmonkey: a PHA-L study of subcortical projections. J Comp Neurol1992; 315: 137–59.

Sandkuhler J, Maisch B, Zimmermann M. The use of localanaesthetic microinjections to identify central pathways: aquantitative evaluation of the time course and extent of the neuronalblock. Exp Brain Res 1987; 68: 168–78.

Sato F, Parent M, Levesque M, Parent A. Axonal branching patternof neurons of the subthalamic nucleus in primates. J Comp Neurol2000; 424: 142–52.

Schaltenbrand G, Wahren W. Atlas for stereotaxy of the humanbrain. 2nd ed. Stuttgart: Thieme; 1977.

Scholz E, Bacher M. Problems in the measurement of parkinsoniantremor. In: Findley LJ, Koller WC, editors. Handbook of tremordisorders. New York: Marcel Dekker; 1995. p. 293–311.

Wichmann T, Bergman H, DeLong MR. The primate subthalamicnucleus. III. Changes in motor behaviour and neuronal activity inthe internal pallidum induced by subthalamic inactivation in theMPTP model of parkinsonism. J Neurophysiol 1994; 72: 521–30.

Wichmann T, Bergman H, Starr PA, Subramanian T, Watts RL,DeLong MR. Comparison of MPTP-induced changes in spontaneousneuronal discharge in the internal pallidal segment and in thesubstantia nigra pars reticulata in primates. Exp Brain Res 1999;125: 397–409.

Yelnik J, Percheron G. Subthalamic neurons in primates: aquantitative and comparative analysis. Neuroscience 1979; 4:1717–43.

Received March 30, 2001. Revised May 28, 2001.Accepted June 8, 2001