Download Temporal and spatial alterations in GPi neuronal encoding might

European Journal of Neuroscience, Vol. 24, pp. 1201–1208, 2006
doi:10.1111/j.1460-9568.2006.04984.x
Temporal and spatial alterations in GPi neuronal
encoding might contribute to slow down movement in
Parkinsonian monkeys
Arthur Leblois,1,2,4 Wassilios Meissner,1,3 Erwan Bezard,1 Bernard Bioulac,1,4 Christian E. Gross1,4 and
Thomas Boraud1,4
1
Basal Gang, Laboratoire de Neurophysiologie, CNRS UMR 5543, Université Victor Segalen, 146 rue Léo Saignat, 33076 Bordeaux
Cedex, France
2
Laboratoire de Neurophysique et Physiologie, CNRS UMR 8119, Université René Descartes, 45 rue des Saints Pères, 75270, Paris
Cedex, France
3
Département de Neurologie, CHU Pellegrin, 1 place Amélie Raba-Léon, 33076 Bordeaux, France
4
Laboratoire franco-israelien de neurophysiologie et neurophysique des systèmes, Université Victor Ségalen, 146 rue Léo Saignat,
33076 Bordeaux Cedex, France
Keywords: basal ganglia, globus pallidus, MPTP-monkey, Parkinson’s disease, single-unit electrophysiology
Abstract
Although widely investigated, the exact relationship between changes in basal ganglia neuronal activity and parkinsonian symptoms
has not yet been deciphered. It has been proposed that bradykinesia (motor slowness) is related either to a modification of the activity
of the globus pallidus internalis (GPi), the main output structure, or to a loss of spatial selectivity of the extrapyramidal motor system.
Here we investigate the relationship between movement initiation and GPi activity in parkinsonian non-human primates. We compare
neuronal encoding of movement in the normal and pathological conditions. After dopamine depletion, we observe an increased
number of neurons responding to movement, with a less specific somato-sensory receptive field and a disruption of the selection
mechanism. Moreover, the temporal order of the response of GPi neurons in parkinsonian animals is reversed. Indeed, whereas
muscle activity and movement are delayed in parkinsonian animals, GPi neuronal responses to movement occur earlier and are
prolonged, compared with normal conditions. Parkinsonian bradykinesia could thus result from an impairment of both temporal and
spatial specificity of the GPi response to movement.
Introduction
Parkinsonian motor symptoms are classically described as (i) tremor,
(ii) rigidity and (iii) akinesia (Agid, 1991). The term akinesia literally
means ‘no movement’ and describes the impairment of voluntary
motor performance. Akinesia encompasses a rich symptomatology
ranging from true akinesia to, more often, several degrees of
bradykinesia, e.g. the delay and slowing down in the execution of a
voluntary movement. Whereas rigidity and tremor can be studied at
rest, the analysis of voluntary movement impairment needs to be
carried out during execution of a motor task.
Basal ganglia (BG) activity is impaired by dopamine depletion in
Parkinson’s disease. The exact role of this network in motor control
remains unclear. As a delay in reaction time is part of parkinsonian
bradykinesia (Benazzouz et al., 1992; Lalonde & Botez-Marquard,
1997), it was initially thought that the BG are involved in the initiation
of movement (Chevalier & Deniau, 1990; Graybiel, 1990; Agid,
1991). This hypothesis is now generally rejected. First, in the normal
animal, neurons of the globus pallidus internalis (GPi), the main
output of the BG, only respond either once movement has begun or
once electromyogram (EMG) activity starts to be modified (DeLong,
Correspondence: Dr A. Leblois, 2CNRS UMR 8119, as above.
E-mail: [email protected].
Received 5 April 2006, revised 30 May 2006, accepted 31 May 2006
1972; Brotchie et al., 1991; Mink & Thach, 1991a; Turner &
Anderson, 1997). Second, the execution of movement is slowed by
lesion and cooling of the GPi, while reaction time is not affected
(Mink & Thach, 1991b; Kato & Kimura, 1992). Therefore, the most
accepted current hypothesis considers the BG to act as an ‘action
selection’ network under dopamine control by disinhibiting desired
cortical motor program generators and by inhibiting other competitive
inappropriate ones (Mink, 1996).
According to this hypothesis, the ability to select one motor
program relies on the precise pattern of response in the BG during the
execution of a motor task. This pattern of response is impaired after
dopamine depletion in Parkinson’s disease, hence the appearance of
major motor symptoms, such as bradykinesia, due to an inability to
select properly one motor program. However, this hypothesis does not
take into account the time dimension underlying motor control (Roux
et al., 2003) in spite of the possible contribution of dopamine in timing
synaptic events through its modulation of GABAergic transmission
(Plenz, 2003). Finally, another hypothesis is that modifications in the
spontaneous firing pattern could be related to parkinsonian symptoms
(Bergman et al., 1994; Wichmann & Delong, 1999).
The present study addresses the issue of information timing in the
BG by analysing extracellular neuronal activity of GPi sensory motor
neurons in three macaque monkeys during voluntary movement before
and after the induction of a moderate parkinsonian syndrome. Our aim
ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd
1202 A. Leblois et al.
Experiments were conducted on three female macaque monkeys
(Macaca mulatta) weighing 3–4 kg (monkeys B, P and J). The
animals were attended to by veterinarians skilled in healthcare and
maintenance of non-human primates. Experiments were carried out in
accordance with the NIH Guide for the care and Use of Laboratory
Animals (1996) and the European Communities Council Directive of
24 November 1986 for care of laboratory animals (86 ⁄ 609 ⁄ EEC).
reward of fruit juice. The recording protocol began once animals
achieved a 90% success rate. During each recording session, at least
15 trials were carried out, during which muscle activity (EMG) was
recorded using intramuscular electrodes acutely inserted into the
deltoid, extensor and flexor carpi radialis, pectoralis, triceps and biceps
muscles. The biceps muscle EMG recordings were chosen for analysis
because this muscle was the first to react during movement (data not
shown). The EMG was amplified and integrated with a time constant
of 20 ms (Benazzouz et al., 1992). The onset of muscle activity is
defined as the last bin in which the myogram increased by 10% for an
epoch that reached a threshold of more than 2.5 times the standard
deviation of the mean. For monkey J, we also recorded the onset of the
movement. Response to passive movement was tested for monkeys B
and P as previously described (Boraud et al., 2000).
Surgical procedure and intoxication
Data analysis
A recording chamber was stereotactically installed under general
anesthesia (ketamine hydrochloride 10–15 mg ⁄ kg i.m., Panpharma,
France, and xylazine 1.5–2.5 mg ⁄ kg i.m., Sigma, St Louis, MO,
USA) at an angle of 45 to the sagittal plane to facilitate the
positioning of the microelectrodes. For the induction of the parkinsonian syndrome, the monkeys were treated daily (17:00 h) with
1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) hydrochloride
(0.2 mg ⁄ kg, i.v., Sigma) dissolved in saline as previously described
(Bezard et al., 2001). The monkeys were assessed daily with the
Benazzouz rating scale (Benazzouz et al., 1995). Injections were
ceased after the animal reached a severity of 5 ⁄ 25 on this scale (i.e. a
cumulative dose of 2 mg ⁄ kg).
Single unit activity was then analysed off-line in relation to passive
(Boraud et al., 2000) or voluntary movement. Mean firing rate and
pattern analysis at rest were performed as previously described
(Boraud et al., 2001). For each trial (i), the neuron firing rate time
course Fi(t) was first determined with a time bin of 10 ms by a
kernel estimator (Silverman, 1986) in which the spike times itT itj
iti
were
convolved
with
a
kernel
function
K(t):
j
Fi(t) ¼ Rnj ¼ 1 K(t ) T i ). We used a gaussian kernel K(t) ¼ exp
(–t2 ⁄ (2s2)) ⁄ (s2P)), where s determined the kernel width, controlling the degree of smoothing. We took s ¼ 0.25 ⁄ F, where F is the
mean firing rate of the neuron over the recording period (Baker &
Gerstein, 2001). The mean firing rate of the neuron across n trials
(n > 15), and aligned on the corresponding stimulus, was then
calculated, to give a smoothed version of the standard peri-event time
histogram (PETH) (Baker & Gerstein, 2001). Smoothed PETHs were
calculated for each recorded neuron, in relation to the four following
events: the GO signal, the onset of muscle activity, the onset of
movement (for monkey J) and, finally, the reward. The mean and
standard deviation (SD) of the mean rate estimate were determined
over a baseline region (during the 500 ms preceding the GO signal).
A neuron was considered responsive if it modified its firing rate
estimate by more than three times the SD of the mean. The onset
time was defined as the first bin of an epoch where the estimate rate
was modified by 10% from the mean in the same direction (i.e.
activation or suppression). When a neuron was responsive to more
than one stimulus, the slope of the first three bins deflecting from the
mean by more than three SD was calculated for each stimulus. The
steepest slope was considered to correspond to the stimulus to which
the neuron was responsive. For this study we considered only
neurons whose response was correlated best to the onset of the
movement or to the EMG for monkey J and to the EMG for
monkeys B and P. Offset time was defined as the last bin of this
epoch in which the mean firing rate had deviated by more than 10%.
Neurons were classified according to the polarity of their response:
either an increase or a decrease in the firing rate. Occasionally a
neuron responded to a single event with a sequence of two or even
three modulations of its firing rate (11.7% in the normal condition
and 10.1% in the MPTP condition). In this case, the response
polarity was classified according to the first modulation.
was to determine the relationship between sensorimotor and temporal
encoding in GPi and parkinsonian motor slowness.
Materials and methods
Animals
Recording procedure
Recordings were performed on the left hemisphere of the three
monkeys. During recording sessions the monkey’s head was immobilized, and 1–4 glass-coated tungsten microelectrodes (impedance
0.2–1 MW at 1000 Hz) confined within a cylindrical guide (2.2 mm
diameter) were advanced into the GP. Each electrode was separately
advanced and optimally placed in the vicinity of cells. GPi neurons
were first identified by on-line analysis of the electrophysiological
signal at rest (see Boraud et al., 1998). The output of each electrode
was amplified with a gain of 5–20k and band-pass-filtered with a 300–
8000-Hz four-pole Butterworth filter. The electrical activity recorded
from each electrode was sorted and classified on-line using a templatematching algorithm (Worgotter et al., 1986), implemented by a PCbased spike sorter (MSD, Alpha-Omega Engineering, Nazareth,
Israel). The spike trains detected by this system, as well as the
behavioral events and other measurements of the monkey’s behavior,
were recorded for off-line analysis.
Behavioral task
In order to investigate neuronal activity related to movement and to
limit responses of cognitive nature, we designed a very simple motor
task. Monkeys were seated in a primate chair in front of a panel with a
single key (Fig. 1). The left arm was immobilized with Velcro straps.
Trials were initiated after the monkeys touched the handle fixed to
their right armrest with their right hand. After a random period (1000–
1500 ms) a green LED turned on as a GO signal., The monkeys had
1000 ms in the normal condition (1500 ms in the parkinsonian
condition) to press the key with a ballistic forward movement of the
right hand. It had to hold this position for 500 ms in order to receive a
Histology (for methodology see Bezard et al., 2001)
All animals were killed by sodium pentobarbital overdose (150 mg/
kg, i.v.), and the brains were removed quickly after death. Each brain
ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd
European Journal of Neuroscience, 24, 1201–1208
Pallidal timing and Parkinsonian bradykinesia 1203
A
B
Hold
Go
1000
to
1500 ms
C
D
Movement
Reward
500 ms
Fig. 1. Behavioural task. (A) Monkeys were seated in a recording chair with their right arm on an armrest. A red LED indicated the hold condition. (B) A green
LED turned on (GO signal) 1000–1500 ms after the hold position was reached. (C) The animal had to press the button within 1000 ms in normal condition (1500 ms
in parkinsonian condition) and (D), to hold the position for 500 ms, in order to receive a reward of fruit juice. Each session consisted of at least 15 successful trials.
was bisected along the midline, and the two hemispheres were
immediately frozen by immersion in isopentane ()45 C) and then
stored at )80 C. Tissue was sectioned coronally at 20 lm in a
cryostat at )17 C, thaw-mounted onto gelatin-subbed slides, dried on
a slide warmer, and stored at )80 C. Histochemical investigation
showed a dramatic decrease in the number of tyrosine hydroxylaseimmunoreactive neurons in the substantia nigra of the three MPTPtreated monkeys (B, )63.7%; P, )69.1%; J, )84.4%), as compared to
control values obtained in four normal animals, matched for age, sex
and weight.
reward delivery) were made using Student’s t-test (P < 0.05).
Latencies for the three animals were pooled after a one-way anova
test showed that they were comparable (P > 0.5). For each monkey,
the coordinates (x, y, z) of all recorded neurons were compared
between the two experimental conditions (normal and MPTP) for each
monkey using a one-way multivariate analysis of variance. For all
three monkeys, there were no significant differences between the two
conditions (P > 0.05).
Results
Statistical analysis
The effect of passive movement on firing rate (activation or
suppression) was pooled between monkeys B and P after a v2 test
showed that distributions were comparable (P > 0.5). The effect of
voluntary movement on firing rate (activation or suppression) was also
pooled between the three monkeys after a v2 test showed that
distributions were comparable (P > 0.5). The normal and MPTP
conditions were then compared for frequency of distribution (v2,
d ¼ 1, P < 0.05 when v2 > 3.841) (Mushiake et al., 1991). Comparisons of the different latencies (latency between stimulus and onset of
muscle activity, onset and offset of muscle activity, stimulus and onset
of the movement, stimulus and onset of GPi activity, muscle activity
and onset of GPi response, onset of movement and onset of GPi
response, onset and offset of GPi response and onset of movement and
Clinical examination
After MPTP injection (see Methods), the three monkeys developed a
moderate akineto-rigid parkinsonian syndrome, which still allowed
them to perform the motor task. This syndrome was characterized by
rigidity, slowness and decrease in the number of voluntary movements, and loss of dexterity, without tremor or postural disorders
(clinical score ⁄ 25: monkey B ¼ 16.8 ± 1.4, P ¼ 15.3 ± 0.9,
J ¼ 19.3 ± 1.3). This parkinsonian syndrome remained stable until
the end of the experiment.
Neuron database
We recorded 144 GPi neurons in the normal condition, 61 in monkey
B over 14 recording days, 50 in monkey P over 9 days and 33 in
ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd
European Journal of Neuroscience, 24, 1201–1208
1204 A. Leblois et al.
Table 1. Summary of individual and mean data
Individual results
Parameter
Monkey B
(n)
Monkey P
(n)
Monkey J
(n)
Mean results
for monkeys
B+P+J
Firing rate at rest
Normal
MPTP
55.8 ± 16.6
76.5 ± 22.3*
(61)
(43)
62.1 ± 16.7
86.6 ± 26.3*
(50)
(55)
56.4 ± 17.9
57.9 ± 24.4
(33)
(20)
57.3 ± 17.3
77.9 ± 26.4*
(144)
(118)
Firing patterns (random ⁄ regular ⁄ bursting)
Normal
29 ⁄ 16 ⁄ 16
MPTP
12 ⁄ 15 ⁄ 16
(61)
(43)
15 ⁄ 14 ⁄ 21
10 ⁄ 16 ⁄ 29
(50)
(55)
12 ⁄ 17 ⁄ 4
2 ⁄ 11 ⁄ 7*
(33)
(20)
56 ⁄ 47 ⁄ 41
24 ⁄ 42 ⁄ 52*
(144)
(118)
Responses with passive movements
Normal response (%)
Response with MPTP (%)
Normal I ⁄ A ratio
I ⁄ A ratio with MPTP
Normal multijoint response (%)
Multijoint responses with MPTP (%)
45.9
69.8*
0.87
0.36*
7.1
56.7*
(28 ⁄ 61)
(30 ⁄ 43)
(13 ⁄ 15)
(8 ⁄ 22)
(2 ⁄ 28)
(17 ⁄ 30)
52.0
64.3
0.62
0*
11.5
30.0
(26 ⁄ 50)
(27 ⁄ 42)
(10 ⁄ 16)
(0 ⁄ 27)
(3 ⁄ 26)
(9 ⁄ 27)
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
–
–
–
–
–
–
48.6
67.1*
0.74
0.16*
9.3
45.6*
(54 ⁄ 111)
(57 ⁄ 85)
(23 ⁄ 31)
(8 ⁄ 49)
(5 ⁄ 54)
(26 ⁄ 57)
Responses with voluntary movements
Normal response (%)
Response with MPTP (%)
Normal I ⁄ A ratio
I ⁄ A ratio with MPTP
Normal Go–Mv delay (ms)
Go–Mv delay with MPTP (ms)
Normal Go–EMG delay(ms)
Go–EMG delay with MPTP (ms)
Normal Mv duration (ms)
Mv duration with MPTP (ms)
Normal EMG duration (ms)
EMG duration with MPTP (ms)
Normal EMG–Mv delay (ms)
EMG–Mv delay with MPTP (ms)
Normal Go–GPi delay (ms)
Go–GPi delay with MPTP (ms)
Normal GPi–EMG delay (ms)
GPi–EMG delay with MPTP (ms)
Normal GPi–Mv delay (ms)
GPi–Mv delay with MPTP (ms)
Normal GPi duration (ms)
GPi duration with MPTP (ms)
44.3
81.4*
0.29
0.06
n.a.
n.a.
157 ± 16
241 ± 27*
n.a.
n.a.
865 ± 13
1654 ± 10*
n.a.
n.a.
148 ± 9
60 ± 12*
)6 ± 18
)235 ± 22*
n.a.
n.a.
219 ± 15
387 ± 8*
(27 ⁄ 61)
(35 ⁄ 43)
(6 ⁄ 21)
(2 ⁄ 33)
–
–
(27)
(35)
–
–
(27)
(35)
–
–
(27)
(35)
(27)
(35)
–
–
(27)
(35)
46.0
58.2
0.53
0.1*
n.a.
n.a.
119 ± 12
320 ± 23*
n.a.
n.a.
831 ± 13
1639 ± 33*
n.a.
n.a.
119 ± 12
137 ± 52
2 ± 28
)173 ± 21*
n.a.
n.a.
220 ± 13
397 ± 29*
(23 ⁄ 50)
(32 ⁄ 55)
(8 ⁄ 15)
(3 ⁄ 29)
–
–
(23)
(32)
–
–
(23)
(32)
–
–
(23)
(32)
(23)
(32)
–
–
(23)
(32)
81.8
60.0
0.93
0.0*
281 ± 15
480 ± 35*
138 ± 15
284 ± 4*
557 ± 13
884 ± 27*
850 ± 30
1605 ± 21*
142 ± 18
202 ± 21*
140 ± 18
81 ± 4*
)7 ± 26
)204 ± 4*
)144 ± 18
)406 ± 27*
219 ± 11
371 ± 3*
(27 ⁄ 33)
(12 ⁄ 20)
(13 ⁄ 14)
(0 ⁄ 12)
(27)
(12)
(27)
(12)
(27)
(12)
(27)
(12)
(27)
(12)
(27)
(12)
(27)
(12)
(27)
(12)
(27)
(12)
53.5
66.9*
0.54
0.07*
281 ± 15
480 ± 35*
139 ± 21
280 ± 48*
557 ± 13
884 ± 27*
849 ± 24
1645 ± 23*
142 ± 18
202 ± 21*
136 ± 17
74 ± 23*
)3 ± 28
)206 ± 34*
)144 ± 18
)406 ± 27*
220 ± 14
384 ± 19*
(77 ⁄ 144)
(79 ⁄ 118)
(27 ⁄ 50)
(5 ⁄ 74)
(27)
(12)
(77)
(79)
(27)
(79)
(77)
(79)
(27)
(12)
(77)
(79)
(77)
(79)
(27)
(12)
(77)
(79)
(n)
Parameters are provided for each monkey (B, P or J) in each of the experimental conditions (normal and MPTP) where available. Mean values of pooled data are also
provided. When appropriate, the number of neurons is provided in parentheses. From top to bottom: firing rate at rest (FR), firing pattern, parameters of the response
to passive movement and of the response to the motor task (voluntary movements). For the response to passive movement, the percentage of neurons responding is
provided along with the ‘inhibited to activated’ ratio (I ⁄ A) and the percentage of neurons responding to the manipulation of more than one joint. For the response to
voluntary movement, the percentage of neurons responding, the I ⁄ A ratio and the latency between the several parameters of the task (mean ± SD) are provided. Mv,
onset of movement; Mv-duration, duration of movement; EMG, onset of muscle response; EMG-duration, duration of this response; GPi, onset of GPi neurons
response; GPi duration, duration of this response. *MPTP vs. normal, P < 0.05.
monkey J over 8 days. One hundred and eighteen neurons were
recorded in the MPTP-treated condition, 43 in monkey B over
17 days, 55 in monkey P over 15 days and 20 in monkey J over
10 days. In monkeys B and P the recordings in MPTP-treated
condition were performed 48 and 42 days after MPTP intoxication,
respectiveley, whereas in monkey J they were performed immediately
after MPTP intoxication. The mean firing rate of GPi neurons was
significantly increased after MPTP-induced dopaminergic depletion
(77.9 ± 17.3 vs. 57.3 ± 17.3 spikes ⁄ s, P < 0.05). The number of
bursting neurons also increased in the pathological condition (28.5%,
41 ⁄ 144 vs. 44.1%, 52 ⁄ 118). Neurons considered were restricted to a
volume of 6 · 3 · 6 mm (sagittal · lateral · depth) for monkey B,
5.5 · 3.5 · 6 mm for monkey P and 7 · 6 · 6 mm for monkey J
inside the GPi nucleus. There were no significant differences in the
ratio of neurons responding to the task between the different locations
of the recordings. Mean and individual numerical data are summarized
in Table 1.
Passive movements
In monkeys B and P the number of GPi neurons responding to passive
movement increased significantly after MPTP treatment (67.1 vs.
48.6%). The ratio of neurons responding to several joint movements
increased also (45.6 vs. 9.3%). These data are consistent with previous
results obtained in fully akinetic monkeys (Boraud et al., 2000).
Motor task
Voluntary movement consisted of the execution of a simple motor task
(Fig. 1). Biceps EMG was monitored during the movement for the
three monkeys. Onset of movement was recorded in monkey J.
Modifications of the response to the movement task
In the normal condition, 53.5% (77 ⁄ 144) of GPi neurons were
responsive to either movement onset or EMG onset (Figs 2 and 3A).
ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd
European Journal of Neuroscience, 24, 1201–1208
Pallidal timing and Parkinsonian bradykinesia 1205
A
sp/s
Normal
100
221102/2_1
50
0
sp/s
-1000
100
Mv
Onset
1000
ms
Mv
Onset
1000
ms
281102/2_4
50
0
-1000
B
sp/s
MPTP
040203/2_13
100
50
0
-1000
C
100
EMG signal (mV)
1000
ms
Mv
Onset
50
MPTP
Normal
0
0
Go signal
400
800
1200
1600
2000
Time (ms)
Fig. 2. PETH of GPi neuronal activity linked to voluntary movement onset (Mv Onset) in monkey J. (A) Example of two GPi neurons in normal condition. Above,
neuron 221102 ⁄ 2–1 responded with an increase (50 trials) and below, neuron 281102 ⁄ 2–4 responded with a decrease in firing rate (62 trials). (B) Example of GPi
neuron activity in MPTP condition (neuron 040203 ⁄ 2–13, 45 trials). For each trial, a green circle on the raster line shows the GO signal, a red circle the onset of the
EMG and a blue one the reward. Note that the reward often occurred after the end of the time window chosen for display. (C) example of biceps EMG aligned on the
GO signal (t ¼ 0) in the normal (black line) and MPTP (grey line) conditions.
ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd
European Journal of Neuroscience, 24, 1201–1208
1206 A. Leblois et al.
A
40
Normal
MPTP
number of neurons
Only 11.7% of the responding neurons (10.1% in the parkinsonian
condition) displayed multiple responses to one of these stimuli. Thus,
all neurons were classified according to the polarity of their earliest
response (see Methods). The mean firing frequency decreased
significantly in 35.1% (27 ⁄ 77) and increased significantly in 64.9%
(50 ⁄ 77) of GPi neurons, resulting in an inhibited ⁄ activated (I ⁄ A) ratio
of 0.54. After MPTP intoxication, the number of GPi neurons
responding to voluntary movement increased significantly to 66.9%
(79 ⁄ 118, Fig. 3A). In addition the I ⁄ A ratio decreased significantly to
0.07 as 6.3% (5 ⁄ 79) of GPi neurons were inhibited whereas 93.7%
(74 ⁄ 79) were activated.
20
10
0
Impairment of temporal encoding
Go
Signal
100
200
300
B
number of neurons
30
20
10
0
-400
-300
-200
-100
EMG
Onset
100
-400
-300
-200
-100
Mv
Onset
C
16
14
number of neurons
In the normal condition the onset of biceps muscle activity (mean ±
SD; Fig. 3C) occurred 139 ± 21 ms after the GO signal and lasted
849 ± 24 ms. The mean delay between the onset of muscle activity
and the onset of movement was 142 ± 18 ms. The mean GPi response
occurred 136 ± 17 ms after the GO signal and lasted 220 ± 14 ms. In
monkey J, the onset of movement occurred 281 ± 15 ms after the GO
signal and the movement lasted 557 ± 13 ms. GPi activity began
3 ± 28 ms before the onset of muscle activity and 144 ± 18 ms before
the onset of actual movement.
After MPTP intoxication (Fig. 3B), both the onset of biceps EMG
activity and movement recorded in monkey J were significantly
delayed to 280 ± 48 ms and 480 ± 35 ms after the GO signal,
respectively, and lasted longer with 1645 ± 23 ms and 884 ± 27 ms
duration, respectively. The mean delay between onset of muscle
activity and movement was significantly increased to 202 ± 21 ms.
Whereas these data confirm previous reports (Benazzouz et al., 1992),
analysis of the response distribution of GPi neurons in relation to the
onset of movement provided unexpected insights. The activity of GPi
neurons was modified significantly earlier after the GO signal
(74 ± 23 ms) and lasted significantly longer (384 ± 19 ms). The
modification of firing activity of GPi neurons thus preceded the onset
of muscle activity and movement onset by 206 ± 34 and 406 ± 27 ms,
respectively. The timing of the GPi response to movement is therefore
altered to a significant extent in the pathological situation.
30
12
10
8
6
4
2
Discussion
It has been proposed that an abnormal firing pattern in the BG could
be the cause of the motor symptoms observed in Parkinson’s disease
(Bergman et al., 1994; Wichmann & DeLong, 1999). In the present
work, we confirmed that the firing pattern of GPi neurons is modified
in the pathological condition. However, the present study is the first to
show that parkinsonian motor slowness is accompanied by modifications in both spatial selectivity of movement-related activity in the GPi
as well as the timing of motor-related responsiveness, which must
contribute to parkinsonian akinesia because of their concomitance.
In the normal animal, the I ⁄ A ratio during voluntary movement was
below 1 (0.54), as observed in other studies (Turner & Anderson,
1997). The GPi is a GABAergic inhibitory output structure of the
basal ganglia that projects to the motor cortex through the motor
thalamic nuclei. An I ⁄ A ratio below 1 has therefore been associated
with the hypothesis that the function of the BG is to select a desired
movement among cortically stored motor programs and to inhibit
other competing programs (Mink, 1996; Boraud et al., 2000, 2002).
The I ⁄ A ratio decreased dramatically to 0.07 following MPTP
intoxication. Although not significant in one animal, the decrease in
the I ⁄ A ratio was strong and consistent among the three animals. It can
be interpreted as the inability for the GPi to disinhibit its thalamo-
0
-500
Time (ms)
Fig. 3. (A) Distribution histogram of time lapse between Go signal and onset
of GPi neuronal response, in normal and MPTP conditions. (B) Distribution histogram of time lapse between onset of the biceps response (EMG) and
onset of GPi neuronal response, in normal and MPTP conditions. (C) Distribution histogram of time lapse between onset of movement (Mv Onset) and
onset of GPi neuronal response, in normal and MPTP conditions. Note the
smaller number of neurons because the onset of movement was recorded only
in monkey J.
cortical targets, instead exerting widespread inhibition. The BG are
thus unable to perform the selection of motor programs correctly. This
hypothesis has already been proposed by our group with regard to the
analysis of passive movement (Boraud et al., 2000). In severely
impaired parkinsonian animals, the number of GPi neurons responsive
to movement increased, as did the number of neurons responding to
the manipulation of several joints, indicating a concomitant loss in
somato-sensory selectivity and the probable recruitment of other GPi
neurons previously involved in different tasks (Filion et al., 1988;
Boraud et al., 2000). In the present study, although our animals were
ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd
European Journal of Neuroscience, 24, 1201–1208
Pallidal timing and Parkinsonian bradykinesia 1207
only moderately parkinsonian, we made very similar observations.
After MPTP-induced dopaminergic depletion, the number of neurons
that responded to the manipulation of more than one joint was
significantly increased from 9.3 to 45.6%. This increase was consistent
in both animals considered but failed to be significant in one of them,
probably owing to a too small sample size. The ratio of neurons involved
in the motor task was significantly increased from 53.3 to 66.9%. This
increase was observed in two of the three animals, but significant only
in one of them. Note, however, that the total number of recorded
neurons was quite low in the only animal displaying a decreased
number of neurons involved in the task. This apparent inconsistency
should thus be interpreted with caution. The present study may include
neuronal activities related to the activation of various muscle groups
and ⁄ or cognitive processes during the execution of the task. It is thus
possible that this global increase in the number of neurons responding
to the task is due to a change in the motor strategy adopted by the
monkey.
Finally, the GPi shifted from a state in which information provided
by the structure was selective, i.e. (i) disinhibition of specific pathway
against a background of tonic inhibition, (ii) selective somato-sensory
encoding and (iii) selective neuronal territory involved in the task to a
physiopathological state with (i) non specific hyper-inhibition during
the movement, (ii) less selective somato-sensory encoding and (iii)
more neuronal territory involved in the task.
The pathological disruption of the temporal pattern of activity in the
GPi during the task constitutes the main interest of the present work.
In the normal animal, GPi neurons modified their firing activity around
the onset of muscular activity. After MPTP intoxication, the
overwhelming majority of GPi neurons began to modify their firing
considerably earlier than the onset of muscle activity. Because the
latter has been shown to begin later in MPTP-treated monkeys than in
normal animals during a motor task (Benazzouz et al., 1992), this
apparent change in the temporal relationship between GPi neurons and
motor outputs may be corollary to the enhancement of inhibition
exerted by the GPi neurons on their target neurons, with a consequent
delay and slowness in motor output. More surprisingly, after dopamine
depletion, GPi neurons began to modify their activity significantly
closer to the GO signal (74 ± 23 vs. 136 ± 17 ms) in two of the three
monkeys, and the duration of the GPi response was longer in the three
monkeys. An earlier response is thus associated with motor slowness,
which appears to be counter-intuitive. However, we have previously
shown similar modifications in the timing of the task-related activity in
the motor cortex, where neurons discharge earlier and longer after MPTP
intoxication than in the normal animal (Doudet et al., 1990). In the
present study it is worth noting that monkey P, in which the GPi response
did not occur significantly closer to the GO signal after MPTP treatment,
had the longest reaction times (time interval between the GO signal and
the EMG onset). By contrast, the delay between GPi response and EMG
onset was similar in the three monkeys in each condition, and if reaction
time may be partially related to other disruption, movement slowness is
related to this premature response. Parkinsonian bradykinesia may thus
be related to a premature activation of the cortex–BG loop during
movement. In parkinsonian patients or MPTP-treated monkeys, the GPi
returns neural information prematurely to cortical motor areas via
thalamic motor nuclei, which might subsequently delay movement
initiation (Benazzouz et al., 1992).
Until now, the most accepted hypothesis considers the BG
selectively to facilitate or inhibit actions generated in the cortex
(Mink, 1996). The late latency of the neuronal response in the GPi
with respect to movement initiation seems to argue against this
hypothesis. However, we showed in a recent theoretical study that
action selection may follow the early non-selective cortical activation
(Leblois et al., 2006). In this framework, the latency of the response
with respect to the onset of the non-selective cortical input is longer in
the GPi than in the cortex, although the BG network plays a crucial
role in action selection. Moreover, we showed that the set of GPi
neurons involved in the disinhibition of the desired motor program
shrinks after dopamine depletion whereas the inhibition of the
remaining of the thalamo-cortical circuits is spread. This spread
inhibition occurs earlier before movement and lasts longer than in the
normal condition, leading to a broad inhibition of cortical motor
generators and to slowing of the intended movements. Bar-Gad &
Bergman (2001) proposed that the role of BG is to reduce the
dimensionality of sparse cortical information. In the specific case of
motor control, the BG would act more as an information selection
structure rather than one involved in action selection. Our findings
show that temporal and spatial alterations in these information
selection processes could be responsible for movement slowness in
MPTP-treated animals.
Acknowledgements
We thank S. Dovero and C. Imbert for their technical assistance. We are
indebted to P. Brown for his helpful comments on the manuscript. J. Simmers
kindly edited the English text. W.M. was a Marie Curie Fellow of the European
Community (HPMF-2001–01300). This work was supported by the CNRS,
Université Victor Segalen Bordeaux 2, the IFR de Neurosciences (INSERM
N8 – CNRS N13) and by CNRS grant CTI01-01.
Abbreviations
BG, basal ganglia; EMG, electromyogram; GPi, globus pallidus pars interna;
I ⁄ A, inhibited to activated ratio; MPTP, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine; PETH, peri-event time histogram.
References
Agid, Y. (1991) Parkinson’s disease: pathophysiology. Lancet, 337, 1321–
1323.
Baker, S.N. & Gerstein, G.L. (2001) Determination of response latency and its
application to normalization of cross-correlation measures. Neural Comput.,
13, 1351–1377.
Bar-Gad, I. & Bergman, H. (2001) Stepping out of the box: information
processing in the neural networks of the basal ganglia. Curr. Opin.
Neurobiol., 11, 689–695.
Benazzouz, A., Boraud, T., Dubédat, P., Boireau, A., Stutzmann, J.M. & Gross,
C. (1995) Riluzole prevents MPTP-induced parkinsonism in the rhesus
monkey: a preliminary study. Eur. J. Pharmacol., 284, 299–307.
Benazzouz, A., Gross, C., Dupont, J. & Bioulac, B. (1992) MPTP induced
hemiparkinsonism in monkeys: behavioral, mechanographic, electromyographic and immunohistochemical studies. Exp. Brain Res., 90, 116–120.
Bergman, H., Wichmann, T., Karmon, B. & DeLong, M.R. (1994) The primate
subthalamic nucleus. II. Neuronal activity in the MPTP model of
parkinsonism. J. Neurophysiol., 72, 507–520.
Bezard, E., Dovero, S., Prunier, C., Ravenscroft, P., Chalon, S., Guilloteau, D.,
Crossman, A.R., Bioulac, B., Brotchie, J.M. & Gross, C.E. (2001)
Relationship between the appearance of symptoms and the level of
nigrostriatal degeneration in a progressive 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-lesioned macaque model of Parkinson’s disease. J. Neurosci.,
21, 6853–6861.
Boraud, T., Bezard, E., Bioulac, B. & Gross, C. (2000) Ratio of inhibited-toactivated pallidal neurons decreases dramatically during passive limb
movement in the MPTP-treated monkey. J. Neurophysiol., 83, 1760–1763.
Boraud, T., Bezard, E., Bioulac, B. & Gross, C.E. (2001) Dopamine agonistinduced dyskinesias are correlated to both firing pattern and frequency
alteration of pallidal neurons in the MPTP-treated monkey. Brain, 124, 546–
557.
Boraud, T., Bezard, E., Bioulac, B. & Gross, C.E. (2002) From single
extracellular unit recording in experimental and human parkinsonism to the
development of a functional concept of the role played by the basal ganglia
in motor control. Prog. Neurobiol., 66, 205–283.
ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd
European Journal of Neuroscience, 24, 1201–1208
1208 A. Leblois et al.
Boraud, T., Bezard, E., Guehl, D., Bioulac, B. & Gross, C. (1998) Effects of LDOPA on neuronal activity of the Globus Pallidus externalis (GPe) and
Globus Pallidus internalis (GPi) in the MPTP-treated monkey. Brain Res.,
787, 157–160.
Brotchie, P., Iansek, R. & Horne, M.K. (1991) Motor function of the monkey
globus pallidus. I. Neuronal discharge and parameters of movement. Brain,
114, 1667–1683.
Chevalier, G. & Deniau, J.M. (1990) Disinhibition as a basic process in the
expression of striatal functions. Trends Neurosci., 13, 277–280.
DeLong, M.R. (1972) Activity of basal ganglia neurons during movement.
Brain Res., 40, 127–135.
Doudet, D.J., Gross, C., Arluison, M. & Bioulac, B. (1990) Modifications
of precentral cortex discharges and EMG activity in monkey with
MPTP-induced lesions of DA nigral neurons. Exp. Brain Res., 80, 177–188.
Filion, M., Tremblay, L. & Bedard, P.J. (1988) Abnormal influences of passive
limb movement on the activity of globus pallidus neurons in Parkinson’s
disease. Brain Res., 444, 165–176.
Graybiel, A.M. (1990) Neurotransmitters and neuromodulators in the basal
ganglia. Trends Neurosci., 13, 244–254.
Kato, M. & Kimura, M. (1992) Effects of reversible blockade of basal ganglia
on a voluntary arm movement. J. Neurophysiol., 68, 1516–1534.
Lalonde, R. & Botez-Marquard, T. (1997) The neurobiological basis of
movement initiation. Rev. Neurosci., 8, 35–54.
Leblois, A., Boraud, T., Meissner, W., Bergman, H. & Hansel, D. (2006)
Competition between feedback loops underlies normal and pathological
dynamics in the basal ganglia. J.Neurosci., 26, 3567–3583.
Mink, J.W. (1996) The basal ganglia: focused selection and inhibition of
competing motor programs. Prog. Neurobiol., 50, 381–425.
Mink, J.W. & Thach, W.T. (1991a) Basal ganglia motor control. II. Late pallidal
timing relative to movement onset and inconsistent pallidal coding of
movement parameters. J. Neurophysiol., 65, 301–329.
Mink, J.W. & Thach, W.T. (1991b) Basal ganglia motor control. III. Pallidal
ablation: normal reaction time, muscle cocontraction, and slow movement.
J. Neurophysiol., 65, 330–351.
Mushiake, H., Inase, M. & Tanji, J. (1991) Neuronal activity in the primate
premotor, supplementary, and precentral motor cortex during visually guided
and internally determined sequential movements. J. Neurophysiol., 66, 705–
718.
Plenz, D. (2003) When inhibition goes incognito: feedback interaction between
spiny projection neurons in striatal function. Trends Neurosci., 26, 436–
443.
Roux, S., Coulmance, M. & Riehle, A. (2003) Context-related representation of
timing processes in monkey motor cortex. Eur J. Neurosci., 18, 1011–1016.
Silverman, B.W. (1986) Density Estimation for Statistic and Data Analysis.
Chapman & Hall, London.
Turner, R.S. & Anderson, M.E. (1997) Pallidal discharge related to the kinematics
of reaching movements in two dimensions. J. Neurophysiol., 77, 1051–1074.
Wichmann, T. & DeLong, M.R. (1999) Oscillations in the basal ganglia. Nature
12, 621–622.
Worgotter, F., Daunicht, W.J. & Eckmiller, R. (1986) An on-line spike form
discriminator for extracellular recordings based on an analog correlation
technique. J. Neurosci. Methods, 17, 141–151.
ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd
European Journal of Neuroscience, 24, 1201–1208