| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Neurophysiol (February 1, 2003). 10.1152/jn.00475.2002
Submitted on Submitted 28 June 2002; accepted in final form 24 September 2002
REPORT
Departments of 1Rehabilitation Medicine, and 2Physiology and Biophysics, 3Graduate Program in Neurobiology and Behavior, and 4Regional Primate Research Center, University of Washington, Seattle, Washington, 98195
 |
ABSTRACT |
|---|
 TOP
 ABSTRACT
 INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
Anderson, Marjorie E., Nadia Postupna, and Mark Ruffo. Effects of High-Frequency Stimulation in the Internal Globus Pallidus on the Activity of Thalamic Neurons in the Awake Monkey. J. Neurophysiol. 89: 1150-1160, 2003. The reduction in symptoms of Parkinson's disease produced by high-frequency stimulation (HFS) in the internal globus pallidus (GPi) has been proposed to be due to stimulus-induced inactivation of pallidal neurons and resulting disinhibition of thalamic neurons. We tested this in awake Macaca fascicularis by stimulating between pairs of electrodes inserted into GPi under electrophysiological control and recording the responses evoked in thalamic neurons. HFS produced a reduction, not an increase, in discharge frequency during the stimulus train in 77% of the responsive thalamic neurons. Only 16% of the responsive cells showed an increase in discharge during stimulation and, for some of these, stimulation at a similar intensity produced contralateral muscle contraction, a probable sign of current spread to the internal capsule. The few thalamic neurons studied during bursting had a reduction in burst frequency and duration during HFS. We conclude that high-frequency stimulation within GPi does not necessarily facilitate thalamic discharge, and it may act, instead, to interrupt abnormal patterns of thalamic discharge associated with parkinsonian symptoms.
| |
INTRODUCTION |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
High-frequency stimulation (HFS)
within the internal globus pallidus (GPi), the subthalamic nucleus
(STN), or the ventrolateral thalamus has become a treatment of choice
for the drug-resistant symptoms of Parkinson's disease (Alegret
et al. 2001
; Ashby et al. 1998; Benabid
et al. 1994; Benazzouz et al. 1993;
Boraud et al. 1996; Brown et al. 1999;
Gross et al. 1997; Krack et al. 1998a,b; Limousin et al. 1995a,b; Pahwa et al.
1997; Pollak et al. 1996; Siegfried and
Lippitz 1994; Valldeoriola et al. 2001, 2002).
Benazzouz et al. (1993) have also reported a reduction
in rigidity and bradykinesia during stimulation of STN in monkeys after
1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) administration, and
the same group (Boraud et al. 1996) has reported that
HFS in GPi also reduces rigidity and bradykinesia in the same monkey
model. This effect is achieved through electrodes implanted in either
GPi or STN that are activated with continuous trains of biphasic
stimuli delivered at frequencies in excess of 100 Hz. The symptomatic
relief during stimulation is reported to approximate or exceed that
provided by pallidotomy, in which a lesion is made in the
posterolateral GPi (Ashby et al. 1998; Brown et
al. 1999). Thus the paradox exists that symptoms are relieved
when neurons are destroyed in GPi or when electrical stimuli
are delivered repeatedly in the same region.
The mechanisms by which HFS in either GPi or STN may lead to
symptomatic relief are unknown. It is hypothesized that either direct
stimulation within GPi or stimulation within the primary known source
of excitatory input to GPi, the STN, reduces the activity of
the GPi inhibitory output neurons and thus increases the activity of
thalamic target neurons by disinhibition. The reduction of activity at
the site of stimulation-GPi or STN-is proposed to occur 1)
because of a depolarization block of voltage-gated currents in neuronal
elements in the stimulated structure (Benazzouz et al.
1995; Beurrier et al. 2001; Burbaud et
al. 1994) or 2) because of activation of presynaptic
axons that have an inhibitory action on the neurons in GPi or STN
(Benazzouz et al. 1995; Boraud et al.
1996; Wu et al. 2001).
To test the hypothesis that HFS in GPi or STN results in a facilitation
of thalamocortical neurons, one could record from neurons in basal
ganglia-receiving areas of the thalamus during stimulation. One study
has attempted to record the effect of stimulation in STN on neurons in
the ventrolateral nucleus of anesthetized rats and has reported
excitation in 16 of 19 thalamic cells studied (Benazzouz et al.
2000). In reality, however, their conclusions were based on the
change in activity after the stimulus train had ended, since
stimulus artifacts obscured the neuronal activity during the stimulus train.
We have recorded the change in activity of thalamic neurons during
trains of high-frequency stimuli applied between microelectrodes inserted under electrophysiological guidance into GPi of monkeys. By
use of a spike-sorting system, we were able to distinguish stimulus
artifacts from thalamic neuronal spikes and record activity during the HFS. Our results indicate that the primary effect
of HFS using this technique is inhibition, not facilitation, of target thalamic neurons. These results have been reported in preliminary form
(Postupna et al. 2001).
| |
METHODS |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
Data were obtained from two juvenile male monkeys (M. fascicularis; initial weight 3.4 and 4.0 kg). All animal procedures were approved by the University of Washington Animal Care Committee and were in accordance with the Guiding Principles in the Care and Use of Animals (American Physiological Society 2000). Animals were on food restriction during the week, with full rations on the weekend. Water was available in their cages at all times. One animal (monkey D) had received a prior infusion of MPTP into the left putamen via a cannula attached to an implanted osmotic minipump, but his symptoms were not severe enough to evaluate their response to HFS. The other animal was intact.
Surgical procedures
Using standard surgical techniques (Anderson
1978), two circular Cilex chambers (Crist, Hagerstown, MD) were
implanted over craniotomies. One, at a 45° angle from the sagittal
plane, allowed access to the GP, and the second, angled from posterior
to anterior at a 30° angle from the coronal plane, allowed access to
the thalamus. Nylon tubes were also implanted anterior and posterior to
the chambers to accommodate stabilization bars.
Stimulation in the GP
The general location of the GP was first mapped with tungsten
electrodes driven by a hydraulic microdrive. The defining
characteristics of GPi were neurons with high-frequency discharge
generally not interrupted by pauses of 0.5 s or longer that were
deep to neurons with similar frequency, but significant pauses. GPi
neurons were sometimes preceded in the track by a short region with
intermediate, steady discharge rates characteristic of border cells
(DeLong 1971).
After mapping, insulated tungsten or platinum-iridium electrodes were
inserted for HFS. Tungsten electrodes (8 mil) were insulated with a
polymide sleeve (Micro-ML tubing) along the shaft and with multiple
layers of epoxylite over the taper. Epoxilite was ground from the tip
and the tips were plated with particulate iron to give a final
impedance of approximately 5 to 20 K
. Platinum-iridium electrodes,
used for one stimulus pair, were glass insulated. Two electrodes,
separated mediolaterally or anterior-posteriorly by 2 mm, were held in
a screw-driven minimicrodrive and lowered in tandem while recording
from each. An implantation depth was chosen at which high-frequency
discharge, typical of GPi activity, could be recorded from each
electrode. These electrodes were then left in place for 4-5 days to
allow stimulation while recording from thalamic neurons at different
sites. Similar electrodes were reinserted, again under
electrophysiological guidance, after a lapse of 
2 days, at the same
or nearby chamber locations.
Trains of constant current biphasic stimuli were applied between the two pallidal electrodes. The standard stimulus protocol used a train of 100 biphasic pulses, 0.2 ms duration for each phase, 120 Hz, applied at a stimulus intensity of 300 µA. In some cases, trains of 10, 500, or 1,000 pulses were applied, lower intensity pulses (50 to 200 µA) were tested, or the duration of each phase of the stimulus pulse was reduced to 0.1 ms.
Stimulus trains were most commonly triggered when the animal positioned its hand in the central hold zone to initiate a trial in the behavioral task (see following text). To assure that changes in activity were not simply due to task conditions, the stimulus trigger was sometimes switched to 200 ms following the GO tone used to trigger movement. If the animal was not working well, stimulus trains could be triggered manually by the experimenters.
As a test of the possibility that stimulus artifacts might be occluding sufficient neuronal spikes to produce the results described below, we sometimes disconnected the stimulus leads from the electrodes and left them hanging nearby, but disconnected. This produced a stimulus artifact, although its configuration and duration were not identical to those seen when the leads were connected. We will refer to this as FAKE stimulation.
Behavioral task
Monkeys were trained to make center-out arm movements across the
surface of a digitizing pad (Turner and Anderson 1997).
A lightweight splint that crossed the wrist and held the pick-up for
the digitizing pad allowed measurement of hand position in x
and y coordinates. Target locations were displayed to the
animal via a one-way mirrorized sheet of plexiglass that displayed
targets on an overhead computer monitor as virtual images in the plane of the digitizing pad. After holding his hand over a central circular target (hold home) zone for a variable period, a tone triggered movement of the right hand to one of the eight peripheral targets that
was illuminated either coincident with the trigger tone (random) or for
a brief period 1 to 3 s prior to the trigger tone (random precued).
Neuronal recording
The activity of neurons in the thalamus was recorded extracellularly using tungsten microelectrodes similar to those described above, but with a slightly smaller shaft diameter (5 mil). Signals were amplified, and all waveforms that passed a threshold were saved digitally (Multichannel Acquisition Processor, Plexon). One or more spikes were discriminated from a single electrode using a dual time/window digital discriminator (RASPUTIN, Plexon), with an additional discriminator used to detect stimulus artifacts. For monkey D, some data were digitized off-line from VCR-based taped data. Isolation of individual spikes and artifact waveforms was reevaluated and corrected off-line using principal component analysis and visualization of the selected waveforms (Off-line Sort Program, Plexon). Figure 1 illustrates action potentials recorded from monkeys G (Fig. 1, A and B) and D (Fig. 1, C and D). Note that both large (Fig. 1B) and small (Fig. 1D) spikes could be detected between artifacts in the stimulus train.
|
Data analysis
Firing patterns and discharge rates during fixed peristimulus
time periods were determined from Plexon-generated files using NEX
software (Plexon). Statistical comparison of mean firing rate during
stimulation to mean firing rate before and after stimulation trains was
done with one-way ANOVA and Tukey posthoc tests (Systat 9). A
P value 
0.005 was considered significant.
Histology
Marking lesions were made at known depths in selected tracks by passing DC current (10-30 µA for 10-30 s) through the microelectrode.
After recording was completed, animals were anesthetized deeply with pentobarbital sodium and killed by transcardiac perfusion with saline followed by fixation with 4% phosphate-buffered paraformaldehyde. Following fixation the brains were blocked, postfixed, cryoprotected with sucrose, frozen, and cut into 50-µm-thick sections. In monkey D, the tissue was cut in the parasagittal plane to facilitate location of thalamic recording tracks. In monkey G, a coronal plane of sectioning was used to facilitate location of GP simulating tracks.
| |
RESULTS |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
Eighty-four cells (55 in monkey D and 29 in monkey G) were studied in the two animals during trains of high-intensity HFS. Based on the histological reconstructions, together with the type of activity recorded at different depths, 73 of these were determined to be in the thalamus at locations shown in Fig. 2.
|
Effects on thalamic neurons of GP-HFS
INHIBITION. A reduction in activity was the predominant stimulus-evoked effect in cells that showed a change in activity during stimulation. Figure 3A shows an example of such a response. In this case, the firing rate dropped abruptly from a mean rate of 18.2 Hz prior to stimulation to a rate of 6.7 Hz during the stimulus train (Fig. 3A). The reduction was sustained throughout the stimulus train, and firing returned to approximately the prestimulus rate when the stimulus train ended. When the pulse duration was reduced from 0.2 to 0.1 ms for each phase, the inhibition disappeared (Fig. 3B). Figure 3C, in which FAKE stimuli were applied to the disconnected stimulus cables, shows that the stimulus-induced inhibition shown in Fig. 3A was not simply due to occlusion of spikes by stimulus artifacts. Although there is a significant (P < 0.001) reduction in detected spike rate during FAKE stimulation (10.7 Hz during stimulation compared with 14.9 Hz during an equivalent time period prior to or 14.7 Hz following stimulation), this reduction (27%) was far less than the 63% reduction elicited by real stimulation with stimulus trains of the same 0.2 ms duration pulses.
|
|
|
|
|
FACILITATION. Only 7 of 43 thalamic cells that responded to the standard stimulus protocol were excited by the stimulation (16%). As shown in the example of Fig. 7A, bipolar stimulation between two microelectrodes usually evoked an initial phasic burst of activity, followed by a more extended facilitation. The stimulus-evoked facilitation of another neuron, illustrated in Fig. 7B, was elicited by stimulation through a concentric bipolar electrode that was too large to allow neuronal recording to guide its placement in GPi. Stimulation with this electrode also produced an initial phasic discharge of the thalamic neuron, followed by a later peak in activity. Stimulation through this electrode at this depth and intensity (300 µA) also evoked movement of the shoulder. Because the stimulus-evoked overt movement was typical of that evoked by stimulation within the internal capsule, it is likely that there was a direct action of the stimulus on the internal capsule. Because the response pattern-initial phasic activation followed by a sustained increase-is similar to that observed for thalamic neurons excited by HFS through microelectrodes positioned in GPi under electrophysiological guidance, it brings up the possibility that facilitation such as that illustrated in Fig. 7A also was due to stimulus spread to the internal capsule.
|
Changes in burst discharges
Three thalamic neurons in monkey D, all recorded after MPTP treatment, were bursting at the time that HFS was tested. In each of these, the number of bursts and the spikes/burst both were reduced during HFS. As illustrated in Fig. 8, bursts of two to five spikes occurred during the period prior to the 100-pulse stimulus train. During the 833-ms stimulation period, bursts occurred less frequently and never had more than two spikes/burst. Single spikes or doublets did occur approximately one or two times during each 833-ms stimulus period, however. After stimulation, longer busts resumed.
|
Special care must be taken to insure that the apparent reduction in bursts was not just a result of occlusion of spikes by the stimulus artifact. (There would be a higher probability that one spike of a clustered high-frequency burst would be occluded by a stimulus artifact than would a single spike.) The bursts shown here, however, were too long to have occlusion with stimulus artifact as the sole explanation for the reduction in burst discharge during stimulation.
Anatomical distribution of stimulation-induced effects
As shown in Fig. 2, almost all of the thalamic neurons studied were in the ventral thalamus, primarily in VA and VLo. Cells whose activity was reduced during stimulation (red squares) generally were intermixed with cells that showed no response (green circles), although there sometimes seemed to be some clustering.
When the activity of multiple neurons was recorded simultaneously through the same electrode, the cells could show the same or different responses to stimulation in GPi. Figure 9A shows the activity of two such nearby neurons, both of which were inhibited during the stimulus train. Figure 9B, on the other hand, shows the activity of one cell that was inhibited and a second cell, recorded at the same time from the same electrode, that was unaffected during the stimulus train. The difference in the effect on these two cells studied with the same stimulus artifact also argues against the possibility that the reduction in activity during the stimulus train is simply due to occlusion of spikes by the stimulus artifact.
|
Anatomical location of stimulating electrodes in GP
Stimulating microelectrodes were positioned under
electrophysiological control. In all cases, both electrodes were
positioned at sites of high-frequency discharge (HFD) activity
characteristic of GP. This was typically 2-4 mm below the first cells
with HFD, usually with significant pauses, and 1 mm above the point
at which the cellular activity decreased, probably indicative of the
internal capsule or the optic tract on the 45° lateral-medial angled
tracks. On occasion, regular, intermediate frequency activity typical
of "border cells" (DeLong 1971) was encountered
above the implantation point at a position consistent with the internal medullary lamina or deep to it at a point consistent with the substantia innominata. Visually responsive fiber-type activity was
occasionally encountered 2.5- to 3-mm deep to the final position of the
stimulating microelectrodes. Because GPe, in which HFD would first have
been encountered, usually has a thickness of <2 mm using this
trajectory, and because the optic tract can be encountered deep to
lateral portions of GPi, we were confident that both electrodes in each
pair were consistently in GPi.
Histological sections of stimulus sites are shown in Fig. 10. In monkey D, sections were cut in the parasagittal plane to optimize identification of the locations at which thalamic neurons were studied. This meant that the sections cut across the trajectory of the pallidal stimulation electrodes. Figure 10, A and B, show electrode penetrations in the GP in two of these parasagittal sections. White arrowheads point to two sites of intense gliosis that would be consistent with positions at which stimuli were applied over the course of the several days that the electrodes were left in place. Figure 10D shows the rostrocaudal and mediolateral arrangement of each of the four electrode pairs, with electrode positions (indicated by matched symbols) separated by 2 mm in the anterior-posterior or medial-lateral plane. In this animal, 1 µl of dextran-rodamine was injected at the position denoted by the open circle in Fig. 10D. This location, which was in the middle of the stimulating electrode positions used to study most thalamic neurons in this animal, is marked by the black arrow in the parasagittal section of Fig. 10B. It is clearly well positioned in GPi.
|
Sections were cut in the coronal plane in monkey G to optimize identification of the site of GPi stimulation (Fig. 10C). For the last experimental session only, glass-covered electrodes were used for the pair of stimulating electrodes. Passage of current between these electrodes caused the glass to shatter for about 5 mm from the tip up the shaft of each electrode. Figure 10C, which shows the resulting lesion, demonstrates that the tips of the electrodes must have been in GPi. These electrodes were pair 1, as illustrated in Fig. 10E. This pair and pair 3, both of which had at least one electrode at the same anterior plane, were at the locations from which most of the thalamic cells were studied in monkey G (Table 1).
The concentric macroelectrode used for stimulation in one session was inserted, again along the 45° angled track, at the location indicated by the filled circle (position 4) in Fig. 10E. This was the electrode from which shoulder movement was produced by stimulation, and the cell whose activity is illustrated in Fig. 7B was excited. Muscle contraction (of the face) was also evoked by stimulation at 300 µA through one of the microelectrode pairs in GPi, the second pair placed at position 3 (triangles). Another neuron, located rostrally in the thalamus, was excited by this stimulus pair. Again, it is possible that this excitation was, in fact, due to spread to thalamic-destined axons in the internal capsule.
Responses of neurons deep to the thalamus
On some occasions in monkey D the recording electrode track extended deep to the thalamus, sometimes into the GP. The two lateral planes of Fig. 2A from this animal show that several of these deep neurons also were inhibited during stimulation (red squares). Two of these cells, in the most lateral plane, had average discharge rates in excess of 40 spikes/s and probably were in the GP. This indicates that stimulation in GPi also did reduce the discharge of some nearby GP somata.
| |
DISCUSSION |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
The major finding of this study is that high-frequency bipolar stimulation through microelectrodes placed in the internal segment of the GP produced inhibition, not facilitation, of most thalamic neurons affected by the stimulus. Only a small fraction of the thalamic neurons with stimulus-evoked responses was facilitated and, in some of those cases, there was evidence that the stimulus position used for those trials actually resulted in stimulus spread to the internal capsule. These findings are in contrast to the predictions that thalamic activity would be disinhibited during HFS in GPi.
Detection of action potentials during the stimulus train
High-intensity stimuli applied in close proximity to a
high-impedance recording electrode produce significant stimulus
artifacts that can preclude examination during the stimulus train. Some prior studies have dealt with this either by reducing stimulus intensity (Dostrovsky et al. 2000) or by using the
discharge immediately following stimulation to imply what happened
during the stimulus train (Benazzouz et al. 1995, 2000;
Beurrier et al. 2001).
In the present study, stimulus artifacts and neuronal action potentials were sorted by a spike-sorting system that allowed us to examine changes in discharge during the high-frequency stimulus train, even with stimulus intensities of 300 µA or greater. Although neuronal spikes that occurred coincident with the stimulus artifact certainly were lost because of amplifier saturation, several findings indicated that occlusion between the stimulus artifact and neuronal spikes could not account for the marked reduction in neuronal activity during the stimulus train. 1) The percentage reduction in discharge during true stimulation usually was much greater than the percentage of the interstimulus interval occupied by the stimulus artifact. 2) When the activity of two thalamic neurons was recorded simultaneously by the same electrode and during the same stimulus train, the activity of one sometimes remained high at the same time that the activity of the other was almost completely suppressed. 3) On occasions in which the leads to the stimulating electrodes were disconnected, but left hanging close enough to the recording electrode to cause a significant stimulus artifact, the measured discharge rates of sorted spikes decreased only slightly and much less than measured during real stimulation. 4) A few neurons did show an increase in discharge, and many showed no change of activity during stimulus trains. Thus we believe that the measured reductions in thalamic neuronal activity during stimulation were real and not just due to occlusion of spikes by stimulus artifacts.
Comparison to other electrophysiological studies of HFS in GPi
No published studies have examined the effect of HFS in GPi on the
activity of thalamic neurons, but some have examined the activity of
nearby neurons in GPi in awake monkeys with MPTP-induced parkinsonian
symptoms or in awake humans with Parkinson's disease. Boraud et
al. (1996) reported that bipolar stimulation through concentric
macroelectrodes inserted stereotaxically into the head of GPi of
monkeys made parkinsonian with MPTP produced a reduction of discharge
of GPi neurons. The reduction was not as complete as we found in the
thalamus, but the pallidal discharge was reduced from the abnormally
high rate after MPTP without stimulation to one that was comparable to
the rate recorded prior to administration of MPTP (approx 80 spikes/s). In the Boraud et al. study, spikes were detected with a
discriminator, using care to be sure that the stimulus artifact did not
result in loss of spikes. Since only the discriminator output was
shown, however, it is difficult to evaluate the relative size or
configuration of the artifacts compared with the spikes. Usually the
stimulus artifact generated by a 90-µs, 350-µA stimulus current
applied within approximately 2-3 mm of the recording site (as
illustrated in their paper) would be large and, when applied at a
frequency of 120 Hz, could produce artifacts that could occlude spikes
and contribute significantly to the apparent reduction of GPi neuronal
discharge rate that they described.
Dostrovsky et al. (2000) and Wu et al.
(2001) also reported that monopolar stimulation in humans
through a microelectrode in GPi produced a reduction in the discharge
of GPi neurons during the stimulus train. In this study, the activity
of GP neurons was recorded within 250-600 µm of a second
microelectrode used for monopolar stimulation. Stimuli were generally
of low intensity (<20 µA) and low frequency (<50 Hz), although
examples were given of profound depression of GPi discharge when
stimuli 300 Hz and 100 µA were applied.
We also found several cells deep to the thalamus that were inhibited by stimulation in GPi, although most did not have the high-frequency discharge typical of GP and may have been more posterior and medial, some in the hypothalamus. Despite the potential decrease in the discharge of somata in GP during stimulation in GPi, however, we still saw profound inhibition of most thalamic neurons during stimulation.
In other neurophysiolgical studies of the effects of HFS, stimuli
have been applied in STN, the primary source of excitatory input to
GPi. Benabid's group (Benazzouz et al. 1995, 2000)
reported that HFS through concentric macroelectrodes in STN of
anesthetized rats reduced the activity of most neurons studied in STN
and in SNr, a target of excitatory axons from STN. They reported that the discharge of thalamic neurons was increased by stimulation, consistent with a decrease in the activity of inhibitory basal ganglia
output from SNr or the entopeduncular nucleus, the rodent homologue of
GPi. The conclusions of this study, however, were based on activity
immediately following cessation of the stimulus train because stimulus
artifacts obscured effects during stimulation. Likewise, the report
that voltage-gated currents in STN neurons studied in the slice are
transiently blocked by HFS in STN was based on the absence of neuronal
activity after the stimulus train (Beurrier et al.
2001).
Possible reasons for the difference between the hypothesized increase and the measured decrease in thalamic discharge during HFS
DIFFERENCE IN STIMULATING ELECTRODE LOCATION.
In humans, HFS is applied through macroelectrode contacts arranged
linearly along an angled anterior-to-posterior trajectory. Stimulation
may be bipolar between two contacts, usually with a center-to-center
separation of 3 mm, or monopolarly versus the pulse generator.
Different stimulation configurations are tested after implantation and
the one that best reduces symptomatology is chosen. The maximum extent
of GPi in humans along the trajectory usually used is 9 mm (see Fig. 1 in Wu et al. 2001). This should be sufficient to insure
that two macroelectrode contacts used for bipolar stimulation could
both be in GPi, but it does not guarantee that they are. In fact,
monopolar stimulation in GPe can produce effects on akinesia that are
opposite to those produced when the electrode is in GPi. Reports from
two groups (Bejjani et al. 1997; Krack et al.
1998a; Yelnik et al. 2000) indicated that
monopolar stimulation in GPe produces an improvement in akinesia but
may also produce dyskinesias, whereas stimulation in GPi worsens akinesia but stops drug-induced dyskinesias. Thus it is possible that
the site of effective stimulation in the current study, which was
confirmed to be in GPi, would have worsened, rather than decreased parkinsonian bradykinesia.
ELEMENTS ACTIVATED BY HFS.
Determination of the elements activated by stimulation within the CNS
is difficult. McIntrye and Grill (2000), who modeled the
effect of mono- and biphasic stimulus pulses of different pulse
durations applied in the vicinity of populations of intermingled model
motoneurons and axons, determined that, for monopolar anodal stimulation, axons had a lower threshold, whereas the threshold was
lower for the initial segment (cells) if monopolar cathodal stimulation
was used. Symmetrical 0.2 ms duration biphasic pulses, such as those
used in the present study (and in most clinical applications), showed
less selectivity than did monophasic ones and excited both cells and
fibers on alternate phases of the stimulus pulse.
), and it is possible that bipolar stimulation between two microelectodes in the current study would more effectively stimulate GPi somata or initial segments, resulting in enhanced inhibition of
thalamic targets, whereas stimulation with macroelectodes used in
humans or in animal studies would primarily excite inhibitory axons
presynaptic to pallidal output neurons (Benazzouz et al. 1995; Boraud et al. 1996; Wu et al.
2001). We did use a concentric macroelectrode on one occasion
to stimulate in GPi, but because of its size, we could not record the
activity of single cells and confirm, by their discharge pattern, that
the tip was in GPi. Stimulation through this electrode at its initial
position did facilitate one thalamic neuron, but, at the standard
stimulus intensity (300 µA), it also produced contraction of
contralateral muscles. We interpreted this as an indication of
stimulation of capsular fibers.
Chronaxy also is shorter for axons than for somata, and it is possible
that the shorter pulse durations (as short as 60 µs) usually used for
HFS in humans would preferentially activate inhibitory axons
presynaptic to GPi output cells, whereas the 200-µs pulses usually
used in the present study would activate the somata of output neurons instead.
If GP-evoked inhibition of thalamic neurons were due to stimulus
configuration, current density, or pulse duration, we would expect
that, if we reduced the stimulus intensity or pulse duration to a point
at which they favored excitation of inhibitory axons presynaptic to
GPi, the effect at the thalamus would switch from inhibition to
facilitation (disinhibition). In fact, this never happened.
OTHER EVIDENCE THAT HFS MAY NOT REDUCE INHIBITORY PALLIDAL OUTPUT.
Several other recent reports have brought into question the conclusion
that HFS in GPi or STN reduces the output of the structure stimulated.
Windels et al. (2000) reported that HFS through
concentric electrodes in the STN of anesthetized rats increased
extracellular glutamate levels in both GP and SNr, which would be
expected if excitatory STN neurons or their axons were activated, not
inhibited. Hashimoto et al. (2001), who stimulated the
STN of MPTP-treated monkeys through macroelectrodes, also found that,
when the electrode was verified to be in STN, stimulation produced an
increase in the mean discharge rate of GPi neurons. In this case, they
also verified that the same stimulation increased purposeful
contralateral limb movements. Furthermore, Baker et al.
(2001), who had the opportunity to record in a human from
thalamic neurons during HFS in GPi, found no consistent changes in
neuronal activity in Vop, the target of pallidal axons, during
stimulation, although reductions in thalamic neuronal activity were
noted after stimulation was stopped. Thus these reports support the
conclusion of the present study that HFS in GPi does not necessarily
lead to decreased inhibitory pallidal output and facilitation of
thalamic targets.
Other potential mechanisms by which HFS in GPi or STN might reduce the symptoms of Parkinson's disease
Hashimoto et al. (2001) also pointed out that
clinically effective HFS in STN of MPTP-treated monkeys produced a more
regular pattern of activity in GPi. We also found that, when a few
thalamic neurons were studied during bursting discharge, HFS in GPi
appeared to reduce the number of bursts and the number of spikes per
burst. Although this was not a situation in which the monkey had
parkinsonian symptoms that could be evaluated, it again brings up the
possibility that the change in thalamic activity that leads to
symptomatology is an abnormal pattern of activity, rather than a
reduction in thalamic activity as a result of an increase in mean
pallidal discharge rate. In fact, abnormal patterns of activity are
common in pallidal neurons in both humans with Parkinson's disease
(Magnin et al. 2000; Zirh et al. 1998)
and in the GPi, SNr, and STN of MPTP-treated monkeys (Bergman et
al. 1994; Wichmann et al. 1999). It may just be
that the beneficial effects of HFS in STN, GPi, or thalamus (and of
pallidotomies) are primarily a consequence of the interruption of
abnormal patterns of activity in corticothalamic circuits.
| |
ACKNOWLEDGMENTS |
|---|
We thank B. Martin for excellent technical assistance and J. Chhatwal for preliminary contributions to this study.
Funding was provided by the National Institutes of Health Grants NS-38228 and RR-00166.
| |
FOOTNOTES |
|---|
Address for reprint requests: M. Anderson, Dept. of Rehabilitation Medicine, Box 356490, Seattle, WA 98195 (E-mail: andermar{at}u.washington.edu).
| |
REFERENCES |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
This article has been cited by other articles:
|
|
 |
|
  W. Xu, G. S. Russo, T. Hashimoto, J. Zhang, and J. L. Vitek Subthalamic Nucleus Stimulation Modulates Thalamic Neuronal Activity J. Neurosci., November 12, 2008; 28(46): 11916 - 11924. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. D. Johnson and C. C. McIntyre Quantifying the Neural Elements Activated and Inhibited by Globus Pallidus Deep Brain Stimulation J Neurophysiol, November 1, 2008; 100(5): 2549 - 2563. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. D. Dorval, G. S. Russo, T. Hashimoto, W. Xu, W. M. Grill, and J. L. Vitek Deep Brain Stimulation Reduces Neuronal Entropy in the MPTP-Primate Model of Parkinson's Disease J Neurophysiol, November 1, 2008; 100(5): 2807 - 2818. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Liu and M. A. Basso Substantia Nigra Stimulation Influences Monkey Superior Colliculus Neuronal Activity Bilaterally J Neurophysiol, August 1, 2008; 100(2): 1098 - 1112. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Guo, J. E. Rubin, C. C. McIntyre, J. L. Vitek, and D. Terman Thalamocortical Relay Fidelity Varies Across Subthalamic Nucleus Deep Brain Stimulation Protocols in a Data-Driven Computational Model J Neurophysiol, March 1, 2008; 99(3): 1477 - 1492. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. K. H. Tang, E. Moro, N. Mahant, W. D. Hutchison, A. E. Lang, A. M. Lozano, and J. O. Dostrovsky Neuronal Firing Rates and Patterns in the Globus Pallidus Internus of Patients With Cervical Dystonia Differ From Those With Parkinson's Disease J Neurophysiol, August 1, 2007; 98(2): 720 - 729. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. A. Basso and P. Liu Context-Dependent Effects of Substantia Nigra Stimulation on Eye Movements J Neurophysiol, June 1, 2007; 97(6): 4129 - 4142. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Herzog, W. Hamel, R. Wenzelburger, M. Potter, M. O. Pinsker, J. Bartussek, A. Morsnowski, F. Steigerwald, G. Deuschl, and J. Volkmann Kinematic analysis of thalamic versus subthalamic neurostimulation in postural and intention tremor Brain, June 1, 2007; 130(6): 1608 - 1625. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Rubin and K. Josic The Firing of an Excitable Neuron in the Presence of Stochastic Trains of Strong Synaptic Inputs Neural Comput., May 1,&nb |
