JOURNALOF NEUROPHYSIOLOGY
Vol. 72, No. 2, August
1994. Printed in U.S.A.
The Primate Subthalamic Nucleus. II. Neuronal Activity in the MPTP
Model of Parkinsonism
H. BERGMAN, T. WICHMANN, B. KARMON,
AND M. R. DELONG
Department of Physiology, The Hebrew University, Hadassah Medical School, Jerusalem, Israel 91010; and Department
of Neurology, Emory University, Atlanta, Georgia 30322
SUMMARY AND CONCLUSIONS
1. The neuronal mechanisms underlying the major motor
signs of Parkinson’s disease were studied in the basal ganglia of
parkinsonian monkeys. Three African green monkeys were sys-
temically treated with 1 -methyl-4-phenyl- 1,2,3,6-tetrahydropyri-
dine (MPTP) until parkinsonian signs, including akinesia, rigid-
ity, and a prominent 4- to ~-HZ tremor, appeared. The activity of
neurons in the subthalamic nucleus (STN) and in the internal
segment of the globus pallidus (GPi) was recorded before (STN, n
= 220 cells; GPi, y2 = 175 cells) and after MPTP treatment (STN,
n = 326 cells; GPi, y1 = 154 cells).
2. In STN the spontaneous firing rate was significantly in-
creased from 19 t_ 10 (SD) spikes/s before to 26 t_ 15 spikes/s
after MPTP treatment. Division of STN neurons recorded after
MPTP treatment into cells with rhythmic bursts of discharge oc-
curring at 4-8 Hz (as defined by autocorrelation analysis) and
neurons without 4- to ~-HZ periodic activity revealed an even
more prominent increase in the firing rate of the 4- to ~-HZ oscilla-
tory neurons.
3. In GPi overall changes in the average firing rate of cells were
inconsistent between different animals and behavioral states. How-
ever, the average firing rate of the subpopulation of neurons with
4- to ~-HZ periodic oscillatory activity after treatment with MPTP
was significantly increased over that of all neurons before MPTP
treatment (from 53 to 76 spikes/s, averaged across monkeys).
4. In the normal state the percentage of neurons with burst
discharges (as defined by autocorrelation analysis) was 69% and
78% in STN and GPi, respectively. After MPTP treatment the
percentage of cells that discharged in bursts was increased to 79%
and 89%, respectively. At the same time the average burst duration
decreased (from 12 1 t 98 to 8 1 t 99 ms in STN and from 2 13 t
120 to 146 t 134 ms in GPi) with no significant change in the
average number of spikes per burst.
5. Periodic oscillatory neuronal activity at low frequency,
highly correlated with tremor, was detected in a large number of
cells in STN and GPi after MPTP treatment (average oscillation
frequency 6.0 and 5.1 Hz, respectively). The autocorrelograms of
spike trains of these neurons confirm that the periodic oscillatory
activity was very stable. The percentage of cells with 4- to ~-HZ
periodic activity significantly increased from 2% to 16% in STN
and from 0.6% to 25% in GPi with the MPTP treatment.
6. Neurons discharging with periodic bursts at frequencies be-
tween 8 and 20 Hz (average 14.4 Hz in STN and 10.5 Hz in GPi)
were also detected more often after MPTP. The autocorrelograms
of these neurons showed a high degree of dampening. The percent-
age of neurons with 8- to 20-Hz oscillations significantly increased
from 0.7% to 10% in STN and from 0.6% to 12% in GPi after
MPTP treatment.
7. The average magnitude and duration of phasic responses
(increases and decrease in firing rate) to the application of flexion
and extension torque pulses to the elbow tended to be increased
after MPTP treatment in both STN and GPi.
8. The duration of extracellularly recorded action potentials
significantly increased with the MPTP treatment (from 0.67 t 0.1
to 0.81 t_ 0.2 ms in STN and from 0.73 t_ 0.1 to 0.8 t 0.1 ms in
GPi).
9. Tyrosine hydroxylase immunohistochemistry revealed that
the striatum as well as STN and GPi of the MPTP-treated mon-
keys were virtually devoid of dopaminergic terminals.
10. These results support a model in which dopaminergic dener-
vation of the basal ganglia (possibly including extrastriatal sites)
leads to increased tonic and phasic activity in STN and GPi. In-
creased inhibition of thalamocortical neurons by GPi output may
eventually result in akinesia and rigidity, whereas periodic oscilla-
tory activity in the cortico-STN-GPi-thalamic circuitry is possibly
involved in the development of parkinsonian tremor.
INTRODUCTION
Parkinson’s disease is characterized by akinesia, rigidity,
and tremor at rest at a frequency of 4-8 Hz. The pathologi-
cal hallmark of this disease is the destruction of dopaminer-
gic nigrostriatal neurons, leading to a decrease in the striatal
dopamine content (for instance, Hornykiewicz and Kish
1987). Primates treated with the neurotoxin
1
-methyl-4-
phenyl-
1,2,3,6-tetrahydropyridine
(MPTP) develop motor
behavioral and pathological changes that are very similar to
Parkinson’s disease (Burns et al. 1983; Jenner et al. 1986;
Langston et al. 1984). However, low-frequency parkinson-
ian tremor is a rather infrequent sign in most primates
other than African green monkeys (Bergman et al. 1990;
Redmond et al. 1985).
Studies of the neuronal activity in the basal ganglia of
MPTP-treated macaques have largely concentrated on
changes in
“tonic” discharge, i.e., changes of the average
firing rates in these structures (Filion and Tremblay 199 1;
Miller and DeLong 1987 ), and the changes in the “phasic”
responses to peripheral inputs (Filion et al. 1988; Miller
and DeLong 1987). In these studies the average firing rate
of neurons in the internal segment of the globus pallidus
(GPi) was increased, whereas the firing rate in the external
segment of the globus pallidus (GPe) was decreased. Re-
sponses of pallidal neurons to limb perturbation were more
frequent and were increased in magnitude and duration.
From these electrophysiological and additional metabolic
(Mitchell et al. 1989) and histochemical (Gerfen et al.
1990) studies a model has been developed (Albin et al.
1989; Bergman et al. 1990; DeLong 1990) in which the
“direct” and the “indirect” pathway between the striatum
and GPi ( Wichmann et al. 1994a) are differentially affected
by the loss of dopamine: the direct pathway becomes un-
0022-3077/94 $3.00 Copyright 0 1994 The American Physiological Society 507
508 H. BERGMAN, T. WICHMANN, B. KARMON, AND M. R. DELONG
deractive, whereas the projection to GPe, the first portion
of the indirect pathway, becomes overactive, leading to re-
duced activity along the inhibitory GPe-GPi and GPe-
subthalamic nucleus (STN) pathways. The changes in the
direct pathway and in the two branches of the indirect path-
way eventually result in overactivity of GPi and its inhibi-
tory projection to the ventrolateral thalamus (VL). Ac-
cording to this model the neuronal activity in STN in par-
kinsonism is released from the normal tonic inhibition
from GPe. The resulting increased activity of STN may
play a pivotal role in the generation of parkinsonism, be-
cause akinesia, rigidity, and tremor are ameliorated by le-
sions ( Aziz et al. 199 1, 1992; Bergman et al. 1990; Guridi et
al. 1993), depolarization block (Benabid et al. 1993; Ben-
azzouz et al. 1993), or pharmacological blockade (Baron et
al. 1992; Signore et al. 1993; Wichmann et al. 1990a,
1994b) of STN.
As yet the neuronal activity in STN in MPTP-treated
monkeys has not been studied in detail. The only previous
study cf the neuronal activity in STN of such monkeys
examined a rather small number of cells and lacked direct
control data from the same animals (Miller and DeLong
1987). The first major aim of the present study was there-
fore to test more thoroughly the prediction that the tonic
firing rate in STN is increased after MPTP treatment.
Although the tonic changes in discharge in the basal gan-
glia of parkinsonian primates serve at least partially to ex-
plain akinesia, bradykinesia, and rigidity, they fail to di-
rectly explain tremor. Previous studies attributed tremor in
Parkinson’s disease to changes in firing of the olivocerebel-
lopontine system, as demonstrated in primates with mid-
brain lesions with or without the injection of the tremoro-
genie drug harmaline (DeMontigny and Lamarre 1973;
Llinas and Volkind 1973), or to tremor-related rhythmic
discharges in VL nucleus. Regarding the latter hypothesis it
has been proposed that increased tonic inhibition of VL by
GPi output predisposes VL neurons to develop overt oscil-
latory firing (Jahnsen and Llinas 1984a,b). In MPTP-
treated macaques, oscillatory activity in the 12- to 14-Hz
range in GPi neurons has been reported by several authors
(Filion and Tremblay 199 1; Miller and DeLong 1987).
However, this was an inconsistent finding, and the rhyth-
mic neuronal bursts were generally not accompanied by
tremor. Recently a mechanism has been proposed (Pare et
al. 1990) that would translate the periodic oscillatory firing
at 12-14 Hz into the 3- to ~-HZ burst discharges that have
been observed during thalamotomy in parkinsonian pa-
tients (Lenz et al. 1988; Ohye et al. 1989).
The relevance of these studies to the understanding of
parkinsonian tremor in humans is limited because the ani-
mal models used generally lack the pathological and clini-
cal features of the human disease. The recently described
African green monkey model of MPTP-induced parkinson-
ism appears to replicate the human disease more closely
than previous models. These primates develop a prominent
3- to ~-HZ tremor (Bergman et al. 1990; Redmond et al.
1985 ), which is similar to that in parkinsonian individuals.
A second major aim of the present study was therefore to
explore the phasic neuronal activity in STN in MPTP-
treated animals of this species.
The STN is the major source of excitatory input to GPi,
probably exerting a strong influence on the activity of this
nucleus (Hamada and DeLong 1992; Hazrati and Parent
1992; Kitai and Kita 1987). To investigate the effects that
altered activity of STN neurons may have on basal ganglia
output, and as a baseline for data obtained after lesioning of
STN (Wichmann et al. 1994b), the third aim of these ex-
periments was to study the tonic and phasic activity of GPi
neurons.
Parts of this study have been reported in abstract form
(Bergman et al. 1992; Wichmann et al. 1990b).
METHODS
Methods that have been described in detail in the companion
paper (Wichmann et al. 1994a) or elsewhere (Bergman et al.
1990) will be described only briefly here.
Animals and behavioral conditioning
The same three juvenile African green monkeys (rnonk~ys A-
C; Cercopithecus aethiops aethiops,
weight 3-5 kg) that were used
in the previous report (Wichmann et al. 1994a) were also used for
the current study. The monkeys were trained in a step tracking
and a torque holding task and then were treated systemically with
MPTP (see below). Recordings from the basal ganglia were
carried out using a parasagittal approach before (Wichmann et al.
1994a) and after MPTP treatment (this report). All experiments
were carried out in compliance with the National Institutes of
Health Guide for Care and Use of Laboratory Animals.
Recording and data acquisition
Spontaneous and task-related neuronal discharge was recorded
from STN and GPi with conventional electrophysiological meth-
ods (as described in Wichmann et al. 1994a). Because after the
MPTP treatment the monkeys would no longer perform the step
tracking movement in the behavioral task and would not cooper-
ate for an adequate somatosensory examination, this report will
focus on the spontaneous activity of neurons as well as on changes
in discharge induced by the application of elbow torque pulses.
Tremor amplitude was measured with an accelerometer (En-
tran Devices, Fairield, NJ) attached to the wrist of the monkeys.
The output from the accelerometer was amplified, low-pass fil-
tered (O-50 Hz), and digitally sampled and stored for off-line
power spectral and cross-correlation analysis.
MPTP treatment
The monkeys were treated systemically with MPTP hydrochlo-
ride, (Aldrich, Milwaukee, WI; 0.3-0.4 mg/kg, im) for a total of
9- 13 injections over a course of 5-l 5 days. Each monkey was
treated with MPTP until it exhibited moderate to severe parkin-
sonian signs. The total amount of MPTP given to each monkey
differed slightly
(monke-v A: 3.4
mg/kg;
monkey B: 3.9
mg/kg;
monkey C: 3.2
mg/kg). During the entire time between the
MPTP treatment and the end of the recording sessions, careful
observation failed to reveal any signs of recovery.
Data analysis
To calculate overall waveforms of action potentials, averages of
256 consecutive spikes from each cell sampled at 20 kHz with 25
sampling points per spike were computed. The signal-to-noise ra-
tio of each cell was calculated as the sum of squares of the means
divided by the sum of the variances of all sampling points. The
polarity of the initial wave of the spike, the duration from the
SUBTHALAMIC NUCLEUS IN PARKINSONIAN MONKEYS
509
0 ’
20 40 50 60 70 80
Discharge rate (spikeslsec)
FIG. 1. Distribution of subthalamic nucleus (STN) firing rates before
(n = 220) (open bars) and after (n = 324) (filled bars) 1-methyl-4-phenyl-
1,2,3,6-tetrahydropyridine (MPTP) treatments. Data was collected during
the pretorque interval and pooled for 3 monkeys.
beginning of the spike to negative and positive peaks, and the
duration to zero crossing points were computed. Spike duration
was defined as the time from the beginning of a spike to the second
crossing of the zero line. Only spikes with an initially negative
wave and a signal-to-noise ratio >5 were included in the analysis.
The spontaneous firing rate of neurons was calculated during
the pretorque period of the torque task (while the monkey was
required to keep the manipulandum within a 1 So window around
the center). Average firing rates were calculated for each individ-
ual monkey, for STN and GPi, and for before and after MPTP
treatment. Two-tailed t tests were used to test the significance of
differences between monkeys in the same state and for individual
monkeys before and after MPTP. Data were pooled for all neu-
rons only if no significant difference (P > 0.05) between monkeys
was found. Relative MPTP-induced changes of the firing rate were
calculated for each animal and averaged for all monkeys.
The firing patterns and correlations of neurons were calculated
during quiet wakefulness in the primate chair between blocks of
trials (“quiet sitting” period). Time interval histograms (TIH),
logarithmic TIH ( LgTIH), and autocorrelation and cross-correla-
tion functions were calculated for lag times of 500 (resolution 1
ms) and 2,500 ms (resolution 5 ms) for all neurons and for all
simultaneously recorded pairs of neurons with acceptable isola-
tion and discrimination quality, as described in the companion
paper (Wichmann et al. 1994a). The histograms were smoothed
by convolution with a normal curve with variable width (Gauss-
ian moving average). This procedure minimizes the distortion of
the histograms in both the time and the frequency domain (Abeles
1982). A normal curve is fitted to the LgTIH and the fit to data,
the modal (most common) interval, and the geometric standard
deviation of the curve (the degree of scatter around the modal
interval) were calculated (Bums and Webb 1976; Paisley and
Summerlee 1984). Confidence lines (0.5% and 99.5%, Abeles
1982) for the auto- and cross-correlation functions were com-
puted with the (null) assumption that the number of spikes in any
bin of the histogram should fit an independent Poisson distribu-
tion (e.g., random firing pattern). Nonrandom discharge of indi-
vidual cells, or nonindependent firing of pairs of neurons, was
assumed if bins in the correlograms were found outside the confi-
dence limits.
A feature-extracting program was used to detect and quantify
features in the correlograms that signified nonrandom discharge
in individual neurons or nonindependent discharge in pairs of
simultaneously recorded cells
(P
< 0.0 1). Autocorrelograms were
classified ( see below) as describing a Poisson process, a burst dis-
charge ( 1 initial peak), or periodic oscillatory functions (equally
distant multiple peaks). The same algorithm was used to detect
and grade periodic oscillatory activity on a scale of 0- 10, using the
jitter of the interpeak intervals (maximally 15% to be considered
oscillatory), the number of significant peaks in the studied func-
tions ( 22 equidistant peaks to be considered oscillatory), the mod-
ulation depths ( oscillation amplitude/ average firing rate), and
the smoothness of the studied function (see details in Karmon and
Bergman 1993). Grade 5 was arbitrarily selected as the minimal
grade for the detection of periodic oscillations. If periodic oscilla-
tory activity was detected in the autocorrelograms, the peaks in the
TIHs were compared with the oscillation frequency. The grades of
oscillatory qualities and the main parameters of periodic activity
(e.g., oscillation frequency, number of periodic cycles, and modu-
lation depth) were stored for later analysis. For correlograms with
one peak (e.g., autocorrelograms of neurons with burst discharge)
the duration and area of the peak ( Abeles 1982) was calculated
and stored in a data base file. Differences in the proportions of
cells in different categories in the normal and in the MPTP-treated
state were statistically evaluated by x2 tests.
Neuronal responses to elbow torque pulses were analyzed as
described in Wichmann et al. ( 1994a). Briefly, the interspike in-
terval data were used to calculate average and SD of discharge
rates across trials for the pre- and posttorque epochs. A given cell
was classified as responding to the torque pulse if its discharge rate
deviated at any time during the posttorque epoch by >2.5 SD
from the mean discharge rate of the pretorque epoch for >20 ms.
Amplitudes and durations of responses were calculated using the
onset and offset times of the responses and the average firing rate
during the pretorque epoch. Only nonoscillatory cells were in-
cluded in the analysis of neuronal responses to torque pulses, be-
cause strong spontaneous oscillations of the neuronal activity pre-
vented a reliable study of torque-related neuronal responses.
TABLE 1.
Spontaneousjring rate of neurons in
STN
before and after
MPTP
treatment
After MPTP
4- to 8-Hz
Before MPTP
Oscillatory neurons Other neurons All neurons
Monkey
A
17.2 & 9.0 (154)
34.9 f 12.4 (40)* 22.5 k 14.0 (171)* 24.8 -t 14.4 (21
l)*
B
21.6
+ 11.8 (12)
38.2 f 8.2 (4)** 35.3 -t 19.8 (23)** 35.7 + 18.1 (27)**
c 22.8 + 12.0 (54) 34.9 k 12.9 (7)** 24.1
+ 13.4 (49)
25.4 + 13.6 (56)
Values are means + SD, with number of neurons in parentheses. Oscillatory neurons and oscillation are defined by a feature-extracting algorithm (see
METHODS). Data collected during the pretorque epoch of the behavioral task. STN, subthalamic nucleus; MPTP, 1 -methyl-4-phenyl- 1,2,3,6-tetrahydro-
pyridine. *
P < 0.0
1; **
P
< 0.05; significant difference from the before MPTP value (2-tailed t test).
510
H. BERGMAN, T. WICHMANN, B. KARMON, AND M. R. DELONG
TABLE
2. Spontaneousjring rate
of
neurons in GPi before and aper MPTP treatment
After MPTP
Before MPTP
4- to 8-Hz
Oscillatory neurons Other neurons All neurons
Monkey
A
B
c
44.2 AI 14.8
(16)
75.4 + 20.4
(15)* 50.1 _+ 26.2 (37)
57.4 f 27.1 (52)
50.5 f 13.2 (21) NA NA 38.8 AI 13.4 (7)
65.0 -t 15.1 (83) 77.1 t 25.6
(1 l)**
62.3 AZ 25.2 (38) 65.6 AI 26.2 (49)
Values are means + SD, with number of neurons in parentheses. Oscillatory neurons and oscillation are defined by a feature-extracting algorithm (see
METHODS). Because the number of neurons recorded in GPi of
monkey B
after treatment with MPTP is small, the data obtained from this monkey and
marked “NA” are not further broken down. GPi, internal segment of globus pallidus. For other abbreviations, see Table 1. *
P <
0.01; **
P -c 0.05;
significant difference from the before MPTP value (2-tailed t test).
RESULTS
Data base
As in the companion paper (Wichmann et al. 1994a),
the data base contained different sets of neurons for differ-
ent parts of the study. Three hemispheres were sampled by
94 penetrations in the normal state, whereas four hemi-
spheres were sampled in 79 penetrations in the post-MPTP
period.
Overall, 220 STN cells from the control state and 294 in
the post-MPTP period were used for the study of neuronal
responses to the application of elbow torque pulses. Firing
pattern analysis was carried out in 291 neurons in the nor-
mal state and 326 STN neurons from the post-MPTP phase
of the experiments. Cross-correlation analysis was carried
out for 72 pairs of cells in STN recorded simultaneously
with a single electrode in the normal state and 25 neuronal
pairs similarly recorded in the post-MPTP state.
In GPi 120 cells from the normal stage and 108 neurons
from the post-MPTP stage were used for the study of neuro-
nal responses to the application of torque pulses. Firing
pattern analysis was carried out in 175 neurons in the nor-
mal state and 154 cells in the post-MPTP state. Cross-corre-
lation analysis was carried out for 18 pairs of GPi neurons
in the normal state and 8 pairs recorded in the post-MPTP
state.
Behavioral eflect of MPTP treatment
and most prominent sign of parkinsonism. During the
course of injections and the first few days thereafter, the
monkeys became more and more akinetic, until they were
eventually unable to feed or groom themselves. Several
days after the development of akinesia, rigidity and tremor
were detected. Low-frequency tremor occurred usually dur-
ing postural adjustments or states of increased arousal of
the monkeys. The amplitude of tremor was often increased
just before the onset of movements, but tremor was not
observed during movements. Only in a few instances, el-
bow torque pulses (0.1 Nm, 60 ms) were effective in reset-
ting the tremor. In the power spectra of accelerometer re-
cordings of tremor, the fundamental frequency was always
<8 Hz (see Bergman et al. 1990 and Fig. 4C).
The parkinsonian symptoms of all three monkeys were
stable for the entire recording period (30, 17, and 27 days,
respectively, after the MPTP treatment). Monkey C, the
most severely affected monkey, died unexpectedly, whereas
the remaining two monkeys were used for the third part of
this series (Wichmann et al. 1994b).
Waveform of extracellularly recorded action potentials
The waveforms of extracellularly recorded action poten-
tials (spikes) in STN and GPi had an initial negative wave
followed by a smaller positive deflection. Neurons with ini-
tial positive spikes, representing - 10% of recorded spikes,
were not included in the analysis.
In both structures the negative peak tended to be larger
The chronic effects of the MPTP treatment have already
than the positive peak. The spike duration was 670 t 100
been described in detail (Bergman et al. 1990). Briefly, (SD) ps in STN (n = 72) and 730 t 130 ps in GPi (n =
after three to eight injections akinesia developed as the first 132). This difference mainly reflected a longer duration of
TABLE
3.
Oscillation frequencies
of
neuronal activity
of
single neurons in the basal ganglia
of
normal and MPTP-treated monkeys
Total Oscillatory Cells
4-8
Hz
8-20 Hz >20
Hz
STN
Before MPTP
After MPTP
GPi
Before MPTP
After MPTP
291
11 (3.8)
6 (2.1) 2 (0.7) 3
(1.0)
326 89* (27.3) 53” (16.3) 34* (10.4) 2 (0.6)
175 9 (5.1) 1 (0.6) 1 (0.6) 7 (4.0)
154
62* (40.3) 39* (25.3) 19*(12.3) 4 (2.6)
Values are numbers of cells, with percentages in parentheses. Neuronal activity was recorded during the quiet sitting state. Oscillatory neurons are
defined by the feature-extracting algorithm (see METHODS). Total, total number of cells. Oscillatory cells, number and percentages of all oscillatory
neurons. 4-8 Hz, number and percentages of cells with 4- to ~-HZ periodic oscillations. 8-20 Hz, number and percentages of cells with 8- to 20-Hz
periodic oscillations. ~20 Hz, number and percentages of cells with oscillation frequency >20 Hz. For abbreviations, see Tables 1 and 2. * Significant
(P
< 0.01) difference between data obtained before and after MPTP treatment.
A STN
40 7
30-
to
=
b-
s
10 -
0 --
I
0
B GPi
40 -
35 -
30-
25 -
cn
a,
0
20
-
s
15 -
10 -
5
0 -i-
I
2
SUBTHALAMIC NUCLEUS IN PARKINSONIAN MONKEYS 511
increased (P < 0.0 1, 2-tailed t test) after MPTP treatment
compared with controls (Fig. 1). No significant changes (P
> 0.05 ) were found between monkeys (Table 1). The aver-
age firing rate in the normal state during the pretorque pe-
riod was 18.8 t 10.3 spikes/s (mean t SD, n = 220) and
25.8 t 14.9 spikes/s (n = 294) after MPTP treatment.
When 4- to ~-HZ oscillatory cells (see below) from the
post-MPTP phase of the experiment were analyzed sepa-
rately, these cells were found to have a much higher dis-
charge rate than all other cells (Table
1).
GPi.
For cells in GPi the average firing rates were quite
variable between individual monkeys. As in STN, the over-
all tonic firing rate of GPi neurons after the MPTP treat-
]flj,J,I,J, ,, I ble2)
ment tended to be increased, but less consistently (Ta-
When 4- to ~-HZ oscillatory cells (see below) cells were
3 4 5 6 7 8 9 10
analyzed separately, however, a more consistent increase in
grades
the average discharge rates of the oscillatory cells in GPi
after MPTP treatment emerged (from 53 to 76 spikes/s,
averaged across monkeys; Table 2).
Spontaneous activity-periodic oscillations
STN.
There were significant changes after MPTP treat-
ment in the firing pattern, most notably the appearance of
periodic oscillatory activity. In the normal state very few
STN neurons demonstrated periodic oscillatory activity,
and if they did, then this activity was relatively weak (grades
~6; Table 3, Fig. 2A). After the MPTP treatment 16.3% of
STN neurons demonstrated strong rhythmic bursting activ-
ity close to the tremor frequency (grades ranging 5 10; Figs.
1 1 I I I I I
3 4 5 6 7 8 9 l0
65
1
grades
91\ n A
cn’
FIG. 2. Distribution of oscillation grades before (open bars) and after
(filled bars) MPTP treatment in STN (A: before MPTP, n = 263; after
MPTP, y1 = 369) and in internal segment of globus pallidus (GPi) (B:
before MPTP, n = 88; after MPTP, n = 154). The oscillation grades were
calculated using a multifactorial system (see METHODS) in which grades
l-5 describe minimal oscillatory activity. Grades 6- 10 describe overt peri-
odic oscillatory activity with increased strength and synchronicity. Only
neurons with oscillation grades >5 were designated as neurons with peri-
odic oscillatory activity in this study.
the positive wave in GPi neurons. The average duration of
the initial negative wave in STN was 280 t 40 ps, whereas
the average duration in GPi was 290 t 50 ps.
After the MPTP treatment there was no significant
change in the amplitude ratio between positive and nega-
tive waves. The spike duration, however, was significantly
(P < 0.00 1,
t
test) increased in STN to 8 10 t 170 ps (n =
54) and in GPi to 800 t 120 pus (n = 75) after MPTP
treatment. In both structures this increase equally affected
the duration of negative and positive wave components.
Spontaneous activity-firing rate
STN.
The spontaneous firing rate of STN neurons during
the pretorque period in the torque task was significantly
O!
I I I 1
’ o! ’
I I I I
time (ms)
500
time (ms) 500
FIG. 3. Examples of autocorrelograms of basal ganglia neurons with
periodic oscillatory activity after MPTP treatment. The autocorrelograms
were calculated up to a lag time of 500 ms (bin = 1 ms) and smoothed by
convolution with a Gaussian curve (g = 1 ms) . The oscillation frequencies
and grades were calculated with the feature-extracting program (see METH-
ODS) . A : STN neuron with periodic oscillatory activity at 5.3 Hz (oscilla-
tion grade = 9).
B:
GPi neuron with periodic oscillatory activity at 4.6 Hz
( oscillation grade = 9). C: STN neuron oscillating at 15.6 Hz ( oscillation
grade = 8 ) . D : GPi neuron oscillating at 10.7 Hz ( oscillation grade = 10 ) .
