therapeutic deep brain stimulation in parkinsonian rats ...core.ac.uk/download/pdf/82056767.pdf ·...
TRANSCRIPT
Neuron
Article
Therapeutic Deep Brain Stimulationin Parkinsonian Rats DirectlyInfluences Motor CortexQian Li,1 Ya Ke,1,* Danny C.W. Chan,1 Zhong-Ming Qian,2 Ken K.L. Yung,3 Ho Ko,1 Gordon W. Arbuthnott,4
and Wing-Ho Yung1,*1School of Biomedical Sciences, Faculty of Medicine, The Chinese University of Hong Kong, Shatin, Hong Kong, China2School of Pharmacy, Fudan University, Shanghai 201203, China3Department of Biology, Hong Kong Baptist University, Kowloon Tong, Hong Kong, China4Brain Mechanisms for Behaviour Unit, Okinawa Institute of Science and Technology Graduate University, Okinawa 904-0495, Japan
*Correspondence: [email protected] (Y.K.), [email protected] (W.-H.Y.)
http://dx.doi.org/10.1016/j.neuron.2012.09.032
SUMMARY
Much recent discussion about the origin of Parkin-sonian symptoms has centered around the ideathat they arise with the increase of beta frequencywaves in the EEG. This activitymay be closely relatedto an oscillation between subthalamic nucleus (STN)and globus pallidus. Since STN is the target of deepbrain stimulation, it had been assumed that its actionis on the nucleus itself. By means of simultaneousrecordings of the firing activities from populationsof neurons and the local field potentials in the motorcortex of freely moving Parkinsonian rats, this studycasts doubt on this assumption. Instead, we foundevidence that the corrective action is upon thecortex, where stochastic antidromic spikes origi-nating from the STN directly modify the firing proba-bility of the corticofugal projection neurons, destroythe dominance of beta rhythm, and thus restoremotor control to the subjects, be they patients orrodents.
INTRODUCTION
Deep brain stimulation (DBS) of the subthalamic nucleus (STN)
is now a recognized therapeutic option for Parkinson’s disease
(PD) (Benabid et al., 1994; Benazzouz et al., 1993; Deuschl
et al., 2006; Obeso et al., 2001). However, the exact mechanism
of action of STN-DBS is still not settled (Chang et al., 2008; Gu-
bellini et al., 2009; Montgomery and Gale, 2008). Early studies
favored an inhibitory action of DBS. Functional inactivation, like
depolarizing block and depletion of transmitters, could produce
a lesion-like effect in the STN (Beurrier et al., 2001; Garcia et al.,
2003; Magarinos-Ascone et al., 2002). Later studies suggested
that DBS may exert an excitatory effect on neural elements
and interferes with the abnormal oscillatory and synchronized
activities that are commonly found in the basal ganglia in Parkin-
sonism (Brown and Eusebio, 2008; Eusebio et al., 2011).
1030 Neuron 76, 1030–1041, December 6, 2012 ª2012 Elsevier Inc.
In principle, DBS can directly activate a wide range of neuronal
elements in STN and the surrounding area, including STN
neuronal soma, axons of passage, and also the terminals of de-
scending fibers from the cortex to STN (Deniau et al., 2010; Lee
et al., 2006; Li et al., 2007; McIntyre and Hahn, 2010; Miocinovic
et al., 2006). By activating both afferent and efferent axons, STN-
DBS can potentially generate widespread and heterogeneous
effects at local and distal sites (Hammond et al., 2008; Hashi-
moto et al., 2003;Maurice et al., 2003). In fact, there are a number
of studies in both human and animals suggesting that STN stim-
ulation can evoke or modulate cortical activities, which may be
beneficial to PD symptoms (Dejean et al., 2009; Fraix et al.,
2008; Gradinaru et al., 2009; Kuriakose et al., 2010; Lehmkuhle
et al., 2009). Thus, early experiments in patients undergoing
implantation of STN electrodes demonstrated cortical evoked
potentials that resembled antidromic activation from the elec-
trodes (Ashby et al., 2001;MacKinnon et al., 2005). Later demon-
stration of resonant antidromic cortical circuit activation as
a consequence of STN-DBS in anaesthetized rats (Li et al.,
2007) was followed by demonstration that high frequency sub-
thalamic stimulation in awake rats could release them from the
akinesia that followed application of dopamine antagonists (De-
jean et al., 2009). These studies along with others in different
situations demonstrated that the subthalamic stimulation also
disrupted the beta rhythms in the cortex in akinetic animals (Fraix
et al., 2008; Kuriakose et al., 2010; Lehmkuhle et al., 2009). Along
the same line of reasoning, recent optogenetic experiments sug-
gested that modifying the activity of STN neurons was less effec-
tive than direct cortical stimulation in reversing the movement
deficits following 6-hydroxydopamine (6-OHDA) lesions in mice
(Gradinaru et al., 2009).
Despite these previous studies that suggest the importance of
antidromically activated responses in the cortex in mediating the
beneficial effect of STN-DBS, elucidating the therapeutic mech-
anism of DBS can only rely on direct recordings of the neural
activities during behaviorally effective DBS in freely moving
animals. In this study, we addressed this question by making
recordings of both single-unit activities and local field potentials
in the motor cortex (MI) of freely moving hemi-Parkinsonian
animals before, during, and after STN-DBS. The results not
only better characterize the abnormal activity in single motor
Neuron
Deep Brain Stimulation Directly Influences Cortex
cortical neurons in Parkinsonism, but also reveal a mechanism
by which STN-DBS directly interferes with the pathological
cortical oscillations characteristic of PD.
RESULTS
Alleviation of Motor Deficits by High FrequencySTN-DBS in Hemi-Parkinsonian RatsWe generated the conventional hemi-Parkinsonian model by
unilateral injection of 6-OHDA into the medial forebrain bundle
(MFB) of the adult rat brain. Successful lesion of the nigrostriatal
pathway was confirmed by the apomorphine-induced contralat-
eral rotation test. Then, a stimulating electrode was targeted at
the ipsilateral STN stereotaxically. In some hemi-Parkinsonian
rats, two 16 channel recording arrays were implanted bilaterally
into layer V of the MI (Figure 1A). In this group of animals, two
stimulating electrodes were implanted in the STN bilaterally to
facilitate the identification of layer V MI neurons in both hemi-
spheres (see Experimental Procedures). After all in vivo experi-
ments, correct placements of the stimulating and recording
electrodes were confirmed histologically (Figures S1A and S1B
available online). The dopamine depletion level induced by the
6-OHDA lesion was further evaluated by the tyrosine hydroxy-
lase (TH) immunostaining of the coronal slices at substantia nigra
and striatum (Figures S1C and S1D). In the substantia nigra pars
compacta (SNc), the nigral dopaminergic neuron loss reached
89.5% ± 3.5% (mean ± SEM, 26 rats). In the striatum, the loss
of TH immunoreactivity was 56.8% ± 7.5% (26 rats).
High frequency stimulation (HFS), which consisted of 125 Hz,
60 ms square pulses at an optimal current (see Experimental
Procedures), improved the mobility of the hemi-Parkinsonian
animals in the open arena (Figure 1B). This effect was confirmed
by assessing several parameters in the open field tests, including
the time and number of episodes spent in mobility and freezing,
the averagemobile speed, as well as the time spent in finemove-
ment. For example, as shown in Figure 1C, while the intact
animals (n = 17) spent 48.3% ± 1.7% of time mobile and
10.3% ± 1.6% of time freezing, the hemi-Parkinsonian rats (n =
26) spent significantly less amount of time moving (16.8% ±
2.2%, p < 0.001), but more time freezing (46.7% ± 2.1%,
p < 0.001). When high frequency (125 Hz) STN-DBS was turned
on, the severity of akinesia was largely, though not completely,
reversed. Thus, during DBS, the lesioned animals (n = 26) spent
40.9% ± 2.0% of time mobile (p < 0.001 compared with DBS off;
p < 0.05 compared with intact) and 17.8% ± 1.4% freezing (p <
0.001 compared with DBS off; p < 0.05 compared with intact).
These beneficial effects disappeared immediately when STN-
DBS was turned off. Bradykinesia symptoms, as reflected by
decreased fine movement and reduced mobile speed, were
also evident in the lesioned animals and were similarly alleviated
during the delivery of STN-DBS (Figure 1D). Furthermore, in the
classical apomorphine-induced contralateral rotation test, STN-
DBS resulted in a modest but statistically significant reduction of
the rotation speed, which was measured as the number of turns
per min (pre-DBS: 19.08 ± 0.61/min; DBS: 16.62 ± 0.62/min; p <
0.01, post-DBS: 18.12 ± 0.73/min; n = 26, Figure 1E).
We also characterized the dependence of the therapeutic
effect of the STN-DBS paradigm on the stimulation frequency
N
and pulse width. As summarized in Figure 1F, at constant stim-
ulus width of 60 ms, low frequency (0.2–10 Hz) STN-DBS failed
to alleviate the motor deficit of the hemi-Parkinsonian animals.
However, when the stimulus frequency was 50 Hz and up to
200 Hz, significant improvement was seen in the percentage
of time spent in motion. Among the four effective stimulation
frequencies tested, namely 50, 125, 200, and 250 Hz, the
optimal frequency was 125 Hz, which is in line with those used
in clinical and experimental studies. The efficacy of the DBS
appeared to be less dependent on pulse width. As shown in
Figure 1G, at a constant stimulation frequency of 125 Hz, signif-
icant therapeutic effects could be achieved at pulse width
ranging from 20 to 80 ms. The falling off of efficacy at 100 ms
suggested that the likely target of the stimulation is fibers rather
than cells.
Stimulation of STN Evokes Antidromic Spikes in the MIin Freely Moving PD RatsWe recorded extracellular neuronal activities from the MI layer V
neurons in both intact and 6-OHDA lesioned rats via multi-
channel recording arrays when the animals were awake and
freely moving. Neuronal activities recorded by each channel
were sorted into single units based on the electrophysiological
characteristics of spike waveforms in the principal component
space (Figure S2A). Two major classes of neuronal unit could
be identified. One type of neuron exhibited a relatively long spike
width (�0.5–0.8 ms) and low spontaneous firing rate (<10 Hz),
which were presumed to be pyramidal, projection neurons
(PNs). Compared with the PNs, presumed interneurons (INs)
held shorter spike width (�0.2–0.5 ms), but higher spontaneous
firing rate (�8–45 Hz). Based on the correlation of the firing rate
and spike width (Figure S2B), these two classes of neurons could
be distinguished unambiguously.
The STN is one of the innervation targets of the long-range cor-
ticofugal axons (Kita and Kita, 2012). As shown in Figure 2A, we
found evidence of direct activation of corticofugal projection
neurons (CxFn) in layer V of MI by identifying stimulation-evoked
antidromic spikes in these neurons during STN-DBS. The identi-
fication was based on the fact that the spike (1) could be de-
tected after the stimulus pulse with a short and fixed latency
and (2) collided with a spontaneous orthodromic spike gener-
ated by the same neuron within a short time window prior to
the electrical stimulus. On average, an antidromic spike, if any,
occurred at a latency of 0.92 ± 0.07 ms (n = 88) in five intact
animals, 1.05 ± 0.04 ms in unlesioned (n = 98), and 0.96 ±
0.08 ms (n = 115) in 6-OHDA-lesioned side of eight hemi-Parkin-
sonian animals and would be eliminated via the collision with
a spontaneous spike occurring within a short interval (<1.0 ms)
before the electrical stimulation in the STN. Furthermore, anti-
dromic spikes could be identified only in the class of neurons
that exhibited a low firing rate and long spike width, reinforcing
the conclusion that these neurons are the CxFn. Those PNs
that did not show antidromic spikes were considered as either
non-CxFn or CxFn that were not antidromically activated under
the experimental condition. The percentages of these neurons,
CxFn, and INs are summarized in Table 1. These data were ob-
tained from experiments in which the stimulation sites were
confined to the lateral STN.
euron 76, 1030–1041, December 6, 2012 ª2012 Elsevier Inc. 1031
*
* * **** *** *** *** *** ***
+
EP
STN
MI MI
Pre DBSDuring DBSPost DBS
SNSTN
MI
StrGP
STN
Intact Hemi-Parkinsonian Intact Hemi-Parkinsonian Intact Hemi-Parkinsonian
0Tim
e sp
ent i
n m
obili
ty (%
)
20
40
60
Tim
e sp
ent i
n fre
ezin
g (%
)
0
20
40
60
0
20
40
60
Tim
e sp
ent i
n fin
e m
ovem
ent
(%)
* *
Aver
aged
mob
ile s
peed
(m
/min
)
0
1
2
3
Mob
ile e
piso
des
(/min
)
NS
*** *** ***
0
2
4
6
8
0
2
4
6
8
10
Free
zing
epi
sode
s (/m
in)
Tim
e sp
ent i
n m
obili
ty(%
)
0
20
40
60
*
***
***
20
40
60
Rot
atio
n sp
eed
(n/m
in)
0
5
10
15
A B
C
E
********
***
*
Pre DBS Post Pre DBS Post Pre DBS Post
Stimulation frequency (Hz) Pulse width (μs) Int
act
Lesio
ned
0.2
1
5
10
50 1
25 2
00 2
50Int
act
Lesio
ned 1
0 2
0 4
0 6
0 8
0 10
0 Pre DBS Post
D
Pre DBS Post Pre DBS Post Pre DBS Post
Tim
e sp
ent i
n m
obili
ty (%
)
F
NS
NS
*** *** *** *** *** ****
** 25
G
Intact Hemi-Parkinsonian Intact Hemi-Parkinsonian Intact Hemi-Parkinsonian
0
20
NS
Neuron
Deep Brain Stimulation Directly Influences Cortex
1032 Neuron 76, 1030–1041, December 6, 2012 ª2012 Elsevier Inc.
Neuron
Deep Brain Stimulation Directly Influences Cortex
The Frequency of Antidromic Spikes Correlateswith Therapeutic Efficacy of STN-DBSSince we could identify the antidromic spikes in CxFn unambig-
uously, we asked if there was any relationship between the anti-
dromic spikes and the therapeutic action of STN-DBS. It has
been pointed out that antidromic cortical excitation may not be
as reliable as generally assumed (Chomiak and Hu, 2007). So
first, we determined the reliability of antidromic spike generation
by examining its success rates at different stimulation frequen-
cies. In a pool of 115 CxFn from the lesioned side of eight
hemi-Parkinsonian rats, the antidromic spike reliably followed
each pulse at a low stimulation frequency. Over 80% of stimuli
were followed by an antidromic spike when the stimulation
frequency was from 0.2 Hz to 10 Hz. However, the reliability of
an antidromic spike following an electrical stimulus decreased
dramatically as the frequency of stimulation was increased,
dropping to 46.8% ± 1.5% at 50 Hz, 26.9% ± 1.1% at 125 Hz,
16.1% ± 0.8% at 200 Hz, and 9.33% ± 0.43% at 250 Hz (Fig-
ure 2B). This decrease in the reliability of antidromic spike
production with increasing stimulation frequency resulted in
the highest frequency of antidromic spikes being produced at
around 125 Hz stimulation rather than other frequencies (Fig-
ure 2C). Interestingly, within the therapeutic window of STN-
DBS, i.e., 50–250 Hz, a positive correlation (R2 = 0.783) between
the frequency of antidromic spikes and the beneficial effect of
STN-DBS was observed (Figure 2D). In addition, we found that
HFS stimulation confined to the medial STN rather than lateral
STN resulted in a lower percentage of cortical neurons exhibiting
antidromic spikes, which was also correlated with less motor
improvement (Figure S3).
As we have seen, at high frequencies of electrical stimulation,
such as 125 Hz, a large percentage of antidromic spikes failed
to be elicited. What impact did this have on the discharge
pattern of the antidromic spikes produced? At a low frequency
of stimulation (10 Hz), due to the high success rate, the spike
density histogram (SDH) of antidromic spikes fitted well with
a Gaussian distribution, indicating a regular pattern (Figure 2E).
However, at 125 Hz STN-DBS (Figure 2F), the success or failure
of the antidromic invasion became unpredictable, resulting in a
Figure 1. STN-DBS Alleviates Parkinsonian Motor Deficits Induced by
(A) Schematic sections showing the placement of the stimulating electrode at t
channel recording arrays at layer V of motor cortex. MI, primary motor cortex;
nucleus; STN, subthalamic nucleus. See also Figure S1.
(B) An example of the locomotor activity of a hemi-Parkinsonian rat before (black
(C) Quantification of the locomotor activities of the hemi-Parkinsonian rats, includ
time spent in freezing and freezing episodes per min (right panels). Severe akinesi
STN-DBS. *p < 0.05; ***p < 0.001; NS, not significant; ANOVA.
(D) Analysis of the time spent in finemovement aswell as the averagemobile speed
were significantly improved by STN-DBS. *p < 0.05; ***p < 0.001, NS: not signific
(E) Contralateral rotation was induced by administration of a low dose of apomorp
rotation during 2 min of 125 Hz STN-DBS reached a statistical significance when
0.01, paired t test.
(F) Based on the time spent in mobility, ANOVA repeated-measures analysis revea
at 50 Hz and a maximum at 125 Hz, with the pulse width set at 60 ms. *p < 0.05,
(G) The efficacy of the STN-DBS paradigm was less dependent on the pulse width
stimulation frequency set at 125 Hz. *p < 0.05, **p < 0.01, ***p < 0.001, compare
Error bars denote SEM.
See also Figures S1, S3, and S4.
N
highly random pattern of SDH that was best fit by the Poisson
distribution. At even higher frequencies of stimulation (i.e.,
200 Hz and 250 Hz), the randomness of the antidromic spikes re-
mained, while the success rate of antidromic invasion decreased
remarkably.
High Frequency STN-DBS Normalizes Firing Rateand Pattern of the CxFnWe then examined the effects of a 6-OHDA lesion and STN stim-
ulation on the firing rates of the layer V CxFn in theMI. To analyze
the firing rate, the antidromic spikes were first removed from the
spike traces (see Experimental Procedures). The average spon-
taneous firing rate of the CxFn was found to be reduced after
6-OHDA treatment (intact: 3.20 ± 0.23 Hz, n = 88, five rats;
lesioned: 2.54 ± 0.17 Hz, n = 115, eight rats, p < 0.05, Figures
3A and 3B). In contrast, in the unlesioned side of theMI, no signif-
icant difference in the CxFn’s mean firing rate was found (3.43 ±
0.26 Hz, n = 98, eight rats, NS compared with intact animals).
During the 2 min of STN-DBS at 125 Hz, a significant increase
in the spontaneous firing of the CxFn in the 6-OHDA-lesioned
hemisphere was observed (3.57 ± 0.19 Hz, n = 115, eight rats,
p < 0.01 compared with DBS off; NS compared with unlesioned
or intact animals). This effect of DBS was absent when the stim-
uluswas delivered at a low frequency of 10 Hz.More importantly,
the 6-OHDA lesion also altered the firing pattern of the CxFn by
increasing episodes of burst firing, as defined by the Legendy
surprise method, which could also be reversed by 125 Hz
STN-DBS, but not at 10 Hz (Figures 3C–3E).
The effects of high frequency STN-DBS on the firing activities
of layer V MI neurons may underlie the motor improvement and
be attributable to the antidromic activation from STN. Since
Degos et al. (2008) showed evidence for a direct STN-cortex
projection, it is important to consider the contribution of ortho-
dromic activation in the MI in mediating the observed behavioral
improvement. However, as shown in Figure S4, unlike the layer V
neurons, 125 Hz STN-DBS did not result in changes in the firing
rates of the layer III/IV neurons, the target of the STN-cortex
orthodromic projection, arguing against a major contribution of
this pathway.
Unilateral 6-OHDA Lesion
he subthalamic nucleus ipsilateral to the 6-OHDA lesion and the bilateral 16-
SN, substantia nigra; Str, striatum; GP, globus pallidus; EP: entopeduncular
line, 1 min), during (red line, 2 min), and after (blue line, 1 min) STN-DBS.
ing the time spent in mobility and mobile episodes per min (left panels) and the
a was induced by 6-OHDA lesion, which was reversed by the delivery of 125 Hz
indicated bradykinesia symptoms in the 6-OHDA-lesioned rats. These deficits
ant, ANOVA.
hine (0.5 mg/kg, subcutaneously). The reduction in the number of contralateral
compared with both pre- (5 min) and post-DBS (5 min) periods. *p < 0.05, **p <
led that the beneficial effect of STN-DBS reached a statistically significant level
**p < 0.01, ***p < 0.001, compared with lesioned condition.
. Significant beneficial effect could be achieved by 20 ms up to 100 ms, with the
d with lesioned condition.
euron 76, 1030–1041, December 6, 2012 ª2012 Elsevier Inc. 1033
Per
cent
age
(%)
1ms100μV
Suc
cess
rate
(%)
0
20
40
100
60
80
Ant
idro
mic
spi
ke fr
eque
ncy
(spi
kes/
s)
0
10
20
30
40
Tim
e sp
ent i
n m
obili
ty (%
)
0
10
20
30
50
40
Stimulation frequency (Hz) Antidromic spike frequency (spikes/s) 0 50 100 150 200 250 0 10 20 30 40
Stimulation reference
0
10
20
30
40
Per
cent
age
(%)
02040
8060
100
Per
cent
age
(%)
0 1 2
0 1 2 4
Stimulation reference
Antidromic spikes
10Hz STN-DBS
125Hz STN-DBS
A B
0 50 100 150 200 250
C D
E
F
3
Spike density histogram (SDH)
Spike density histogram (SDH)
R2=0.783
5
0
10
20
30
40
0 1 2 43 5
CxFn-1
CxFn-2
100ms
Antidromic spikes
CxFn-1
Per
cent
age
(%)
0
20
40
80
60
0 1 2
100ms
CxFn-2
Stimulation frequency (Hz)
Figure 2. Identification and Characterization of the Antidromic
Spikes in the CxFn during STN-DBS
(A) An example of the identification of antidromic spikes. Shown on the top are
the spike templates of two single units picked up by the same recording
electrode and discriminated by their spike waveforms. Five overlaid traces are
shown for each unit. An antidromic spike appeared following the stimulation
artifact (arrows) with a short, fixed latency (i.e., 0.92 ms, top left). Collision
occurred when a neuron discharged spontaneously immediately before a
Table 1. Distribution of Interneurons and Projection Neurons
with and without Antidromic Spikes Recorded in MI
Interneurons
(%)
Projection
Neurons without
Antidromic
Spikes (%)
Projection
Neurons with
Antidromic
Spikes (CxFn; %)
Intact (five rats) 52 (24.1%) 76 (35.2%) 88 (40.7%)
Unlesioned
(eight rats)
60 (22.0%) 115 (42.1%) 98 (35.9%)
6-OHDA-lesioned
(eight rats)
71 (24.6%) 103 (35.6%) 115 (39.8%)
See also Figure S2.
Neuron
Deep Brain Stimulation Directly Influences Cortex
1034 Neuron 76, 1030–1041, December 6, 2012 ª2012 Elsevier Inc.
High Frequency STN-DBS Disrupts PathologicalOscillations and Synchrony of MI Neuronal PopulationsApart from altering the firing rate and pattern of individual CxFn,
dopamine depletion induced pathological activities in the MI at
the population level. Figure 4A shows typical raster plots of 10
CxFn after the 6-OHDA lesion, showing clear synchrony among
most of the neurons, which was confirmed by the cross-correla-
tion analysis between pairs of neurons (Figure 4B). Interestingly,
the degree of correlation was dramatically reduced when the
burst firing was filtered out from analysis, indicating that the syn-
chrony was a reflection of the abnormal burst firing. The pooled
data from the recordings in five intact and eight hemi-PD rats
confirmed this observation (Figure 4B). Power spectrum analysis
of the local field potential (LFP) revealed the appearance of oscil-
latory activity at the beta band (�20–30 Hz) after dopamine
depletion (Figures 4C and 4D). STN-DBS at 125 Hz was highly
effective in removing the abnormal synchrony among the neu-
rons and the beta oscillations (Figures 4B and 4D).
The strength of coupling between spike times and LFP at any
given frequency, known as spike-field coherence, was also
investigated. 6-OHDA lesion caused the coherence value to
stimulus pulse (<1.0 ms) and prevented the propagation of the antidromic
response being recorded (top right). However, collision failed to occur if
a longer time (i.e., >1.7 ms) elapsed between the spontaneous spike and the
stimulus pulse (bottom left). Note that the antidromic spike could not collide
with a spontaneous spike generated by another neuron, even within a short
time lag (bottom right). See also Figure S2.
(B) The success rate of the antidromic spike decreased with the increased
frequencies. Data shown here were obtained from 115 CxFn of eight hemi-
Parkinsonian rats.
(C) The absolute frequency of antidromic spikes increased with the stimulation
frequency, reaching a peak that corresponded to 125 Hz stimulation
frequency. Data were from 115CxFn in eight hemi-Parkinsonian rats. In (B) and
(C), open circles represent data from an individual rat; closed circles represent
overall mean.
(D) Within the therapeutic window of STN-DBS frequency (from 50 to 250 Hz,
gray shade), there was a positive correlation between the antidromic spike
frequency and the time spent by the animal in mobility.
(E) At a low stimulation frequency of 10 Hz, highly regular antidromic spikes
were produced in two CxFn. The time of occurrence of the stimulus and
evoked antidromic spikes are shown on the left, and the resulting SDHs, fit by
Gaussian distribution, are shown on the right.
(F) In contrast, 125 Hz STN-DBS generated two distinct discharge patterns of
the antidromic spikes that were highly random and characterized by Poisson-
like SDHs.
See also Figures S2 and S3.
Firin
g ra
te (s
pike
s/s)
Unlesioned side, 125Hz STN-DBS
Lesioned side, 125Hz STN-DBS
Lesioned side, 10Hz STN-DBS
0 1 2 3 4 5 6 05
1015 1
Unlesio
ned
Intact Hemi-Parkinsonian
4
3
2
0
Mea
n fir
ing
rate
(spi
kes/
s)
Time (min)
05
1015
05
1015
% o
f tot
al n
umbe
r of s
pike
s in
bur
st d
isch
arge
% o
f tim
e sp
ent i
n bu
rst d
isch
arge
0
5
10
15
25
20
1
2
3
4
Intact Hemi-Parkinsonian Intact Hemi-Parkinsonian
A B
NS **
*NS
NS
C
NS
*
**NS
NSD E
NS
***
******
NS
Intact
SPKC06bSPKC08c
SPKC10bSPKC11a
SPKC08d
6-OHDA lesioned
125Hz STN-DBS
1s
SPKC08aSPKC09aSPKC09bSPKC15aSPKC16a
SPKC06bSPKC08c
SPKC10bSPKC11a
SPKC08d
10Hz STN-DBSSPKC06bSPKC08c
SPKC10bSPKC11a
SPKC08d
Lesio
ned
125H
z10
Hz
Unlesio
ned
Lesio
ned
125H
z10
Hz
Unlesio
ned
Lesio
ned
125H
z10
Hz
Figure 3. High-Frequency STN-DBS Normalizes Firing Rate andPatterns of CxFn
(A) Typical examples of spike rate histogram, showing the effect of STN-DBS
on the spontaneous firing rates of CxFn. Two minutes of STN-DBS at 125 Hz
(gray shade) caused no effect on the CxFn in the unlesioned side, but
increased the firing rate of the neurons in the lesioned side. However, STN-
DBS delivered at 10 Hz was ineffective in altering the firing rate of the same
CxFn in the lesioned side.
(B) Statistical analyses revealed that the mean firing rate of the CxFn after the
6-OHDA lesion was reduced significantly, but was normalized by 125 Hz STN-
DBS. In these analyses, the antidromic spikes were removed and did not
contribute to the counts. Data were from 88 neurons in five intact rats and 98
neurons in eight hemi-Parkinsonian rats. *p < 0.05, **p < 0.01 ANOVA.
Neuron
Deep Brain Stimulation Directly Influences Cortex
N
increase specifically at the beta band (20–30 Hz; Figures 5A and
5B). Interestingly, the delivery of 125Hz STN-DBS eliminated this
beta band spike-field coherence. In addition, the phase synchro-
nization (or coherence phase) between spikes and LFP at the
beta band was derived and presented as polar histogram (Fig-
ure 5C). It was clear that the 6-OHDA lesion induced a specific
phase-locking phenomenon between spikes and LFP in beta
band (91/115 pairs, eight rats), whereas the pairs from the intact
and unlesioned side showed randomly distributed phases,
ranging from 0� to 360� (intact: 88 pairs from five rats; unle-
sioned: 98 pairs from eight rats). During the delivery of 125 Hz
STN-DBS, the specific phase-locking bias present in the 6-
OHDA-lesioned condition was not notable.
Antidromic Spikes Directly Modulate the FiringProbability of CxFnTo elucidate the possible mechanism underlying antidromic
spikes-mediated beneficial effect, we next asked the crucial
question of whether antidromic spikes can directly modulate
the firing properties of the CxFn. Since the DBS was delivered
at a high frequency of 125 Hz, i.e., with an interstimulus interval
of 8 ms, we analyzed the impact of STN-DBS by selecting and
then aligning all those 8 ms time segments that contained anti-
dromic spikes. A typical example is shown in Figure 6A (left
panel). These time-aligned segments of neuronal activities re-
vealed that the probability of firing of the CxFn was influenced
by the 125 Hz HFS: a complete cessation of firing immediately
following the antidromic spike that lasted for about 1ms followed
by an elevated firing probability in the subsequent 2 ms interval.
The early depression of firing probably represents the refractory
period of the antidromic spikes, while the delayed increase in
firing probability could reflect a change in the intrinsic excitability
of the neurons or functional connectivity within the local circuit as
a consequence of antidromic spikes. In contrast, for those 8 ms
segments that did not contain antidromic spikes, there was
no change in the firing probability (Figure 6A, right panel).
In the unlesioned side, no change in firing probability was
found, whether DBS was turned on or off (Figure 6B). These
results demonstrate that the antidromic spikes directly and
selectively alter the firing probability of the layer V CxFn. On
the other hand, when DBS was delivered at 10 Hz, in addition
(C) Typical examples of raster plots showing the neuronal discharge patterns
under intact, 6-OHDA-lesioned, 125 Hz, and 10 Hz STN-DBS conditions.
Random discharge patterns of CxFn were found in intact rats. In contrast, the
firing of the CxFn in the 6-OHDA-lesioned rat was characterized by increased
burst firing. The burst discharges, as defined by the Legendy surprise method,
were highlighted with red lines. The burst firings of these neurons were clearly
reduced under 125 Hz STN-DBS. In contrast, clear burst firing remained in all
the neurons at 10 Hz STN-DBS.
(D and E) Both the analyses of the % of total number of spikes in burst
discharge (D) and the % of time spent in burst discharge (E) were significantly
increased after 6-OHDA lesion and were significantly reduced by 125 Hz
STN-DBS. *p < 0.05, **p < 0.01, ***p < 0.001; NS, not significant; ANOVA. The
pooled data were obtained from 88 CxFn of five intact rats and 98 and
115 CxFn from unlesioned and lesioned sides, respectively, of eight hemi-
Parkinsonian rats.
Error bars denote SEM.
See also Figure S2.
euron 76, 1030–1041, December 6, 2012 ª2012 Elsevier Inc. 1035
-0.5 0.5 0 -0.5 0.5 0
-0.5 0.5 00
0.1
0.2
0.3
-0.5 0.5 0
Pow
er s
pect
ral
dens
ity (%
)
0
2
4
0 20 40 60Frequency (Hz)
1s200μV
Intact
6-OHDA lesioned
125Hz STN-DBS
Unlesioned
08b vs.12a
Intact Unlesioned 6-OHDA lesioned 125Hz STN-DBS6
B
C
0.4
0
0.1
0.2
0.3
0.4
00.10.20.30.4
00.10.20.30.4
D
ACh01cCh02bCh03aCh05cCh08cCh09cCh11b
Ch14aCh16c
Ch13a
1s
6-OHDA lesioned (-burst) 125Hz STN-DBS
02b vs.14a(-burst)
01c
02b
03a
05c
08c
09c
11b
14a
16c
13a
01c02b03a05c08c09c11b
14a16c
13a
intact 6-OHDA lesioned
06d08b10a11b12a13c14a
16b15b
16d
15b
14a
13b
12a
11c
10a
06b
08b
02b vs.14a
02b vs.14a
01c
02b
03a
05c
08c
09c
11b
14a
16c
13a
01c02b03a05c08c09c11b
14a16c
13a
01c
02b
03a
05c
08c
09c
11b
14a
16c
13a
01c02b03a05c08c09c11b
14a16c
13a
0 0.2 0.4 0.6 0.8 1.0
Lesioned -burst 125Hz
Intact Hemi-Parkinsonian
0.3
0.2
0.1
0
80 1000
2
4
6
0
2
4
6
0
2
4
6
*** ***
NS
0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100
Cor
rela
tion
inde
x
***
Correlation index
Figure 4. High Frequency STN-DBS Ameliorates Abnormal Popula-
tion Responses in MI
(A) Typical raster plots showing the synchronized spikes, especially during
burst discharge, across 10 CxFn from the 6-OHDA-lesioned condition.
(B) Analysis of cross-correlation between pairs of CxFn in 6-OHDA lesion.
The color-coded correlation matrix on the left side of each panel revealed
Neuron
Deep Brain Stimulation Directly Influences Cortex
1036 Neuron 76, 1030–1041, December 6, 2012 ª2012 Elsevier Inc.
to the biphasic changes in firing probability immediately fol-
lowing the antidromic spikes, a slight increase in firing rate at
a much delayed time of around 40–50 ms poststimulation was
observed (Figure S3). This was likely the effect relayed to the
cortex via the basal ganglia circuit under STN-DBS.
DISCUSSION
Previous studies demonstrate that STN-DBS can modulate
activities of the cortical motor areas in both PD patients (Cunic
et al., 2002; Dauper et al., 2002; Kuriakose et al., 2010; Limousin
et al., 1997) and in animal models of Parkinsonism (Dejean et al.,
2009; Lehmkuhle et al., 2009; Li et al., 2007). In this study,
making use of multichannel recording arrays implanted into the
MI, we recorded and analyzed single-unit neuronal activities
from populations of the layer V CxFn of freely moving hemi-
Parkinsonian rats during a therapeutically effective STN-DBS
paradigm. This approach allowed us to directly address several
key questions on the involvement of MI in STN-DBS and
provided insight into a mechanism of the therapeutic action
of DBS.
Despite the fact that MI is a major target of the basal ganglia
output and therefore likely transforms patterns of pathological
activities into motor symptoms, there were only very few studies
on characterizing the firing rate and patterns of primary motor
cortical neurons in Parkinsonism at the single cell level (Goldberg
et al., 2002; Pasquereau and Turner, 2011). In fact, single-unit
activities from large populations of CxFn in freely moving PD
rats in the resting state and during STN-DBS had not been
achieved before. Our findings showed that there were dramatic
changes in the neuronal activities of CxFn at both single-cell
and the population level. The increased burst discharge and
oscillatory rhythm at the beta range are similar to the hallmark
events found in human and animal models of PD (Wichmann
and Dostrovsky, 2011) and in line with previous studies on
Parkinsonian primates (Goldberg et al., 2004; Pasquereau and
Turner, 2011) and rodents (Sharott et al., 2005). The origin of
these changes in the motor cortex, like that in the basal ganglia
increased correlation in many pairs of CxFn in 6-OHDA, but was absent in
intact animals. The correlations were markedly reduced when the burst
discharges were filtered out artificially from the record. The strong correlation
was abolished when high frequency STN-DBS was delivered. The correlation
levels from one pair of neurons in the intact animal and from one pair of
neurons in a 6-OHDA rat (indicated by the yellow squares) are shown on the
right. Statistical comparison between five intact rats and eight hemi-Parkin-
sonian rats is shown in the lower panel. ***p < 0.001; NS, not significant;
ANOVA.
(C) Typical raw LFP waveforms recorded in layer V of MI from both intact and
hemi-Parkinsonian rats. For each hemi-Parkinsonian rat, recordings were
made from both unlesioned and 6-OHDA-lesioned sides and also 125 Hz
STN-DBS on the lesioned side.
(D) Power spectral distribution of MI-LFPs under the above conditions. A clear
increase of power in the beta range (20–30 Hz) was found in theMI ipsilateral to
6-OHDA lesion, which was absent under both intact and unlesioned condi-
tions. STN-DBS of 125 Hz alleviated the dominant beta band oscillatory
activities. Data are averages obtained from 61 LFP records in five intact rats
and 95 LFP records in eight lesioned rats.
Error bars denote SEM.
See also Figure S2.
90°
LFP
Intact 6-OHDA lesioned
400μV
2s
Coh
eren
ce v
alue
00.020.040.06
0 20 40 60 80 100
0.080.1
00.020.040.06
0 20 40 60 80 100
0.080.1
00.020.040.06
0 20 40 60 80 100
0.080.1
00.020.040.06
0 20 40 60 80 100
0.080.1
Coh
eren
ce p
hase
Spike
0°
30°60°90°120°
150°
180°
210°240° 270° 300°
330°
8 6 4
2
8 6 4
2
1510
5
6 4
2
Frequency (Hz)
A
B
C
Intact Unlesioned 6-OHDA lesioned 125Hz STN-DBS
Intact Unlesioned 6-OHDA lesioned 125Hz STN-DBS
0°
30°60°90°120°
150°
180°
210°240° 270° 300°
330°
0°
30°60°90°120°
150°
180°
210°240° 270° 300°
330°
0°
30°60°120°
150°
180°
210°240° 270° 300°
330°
Figure 5. High Frequency STN-DBS
Disrupts the Beta Band Spike Field Coher-
ence in MI
(A) Typical LFP and corresponding spike trains
from CxFn recorded simultaneously from intact
(left) and 6-OHDA-lesioned (right) hemispheres
of a hemi-Parkinsonian rat.
(B) Analysis of the spike-field coherence spectrum
revealed a prominent peak in beta band (20–
30 Hz), which was abolished when high frequency
STN-DBS was delivered.
(C) Representative results of polar-histogram of
the spike-field coherence phase distribution in
beta band (20–30 Hz). Phase-locking bias
between spikes and LFP was evident after
6-OHDA lesion, which was absent in both the
intact rat and unlesioned side of hemi-Parkinso-
nian rats, and reversed to random distribution
when 125 Hz STN-DBS was delivered.
See also Figure S2.
Neuron
Deep Brain Stimulation Directly Influences Cortex
circuit, remains unknown. However, as the output station of the
motor system, these pathological changes in the CxFn likely
contribute to the symptoms in PD. For example, the pathological
enhancement in beta oscillatory rhythm may underlie abnormal
persistence of the status quo and deterioration of behavioral
control (Engel and Fries, 2010). Furthermore, multiple studies
have shown that a critical effect of STN-DBS is the reduction
of the synchronization of oscillatory activities between the basal
ganglia and cortex (Eusebio et al., 2011; Hammond et al., 2007).
Our data confirmed that the spike firing activities of the CxFn
were highly synchronized under PD condition, especially during
burst discharge. These abnormal firing patterns across the
CxFnwere effectively eliminated by STN-DBS. The simultaneous
recording of spikes and LFP by the same recording channel
also allowed us to study the coherence between them. The
results showed that there was increased coherence level
between spikes and local field potentials at beta band only in
the 6-OHDA-lesioned hemisphere. The significance of the in-
creased coherence is unclear, but may contribute to bradykine-
sia and other movement suppression (Brown and Williams,
2005). Beta-band spike-field coherence may also represent
excessive ‘‘stop’’ signals that underlie akinesia in PD (Swann
et al., 2011).
In this study, we also provided direct evidence of the occur-
rence of antidromic spikes in MI during the DBS paradigm.
This finding supports previous studies using electroencephalo-
gram (EEG) recordings that STN-DBS results in the antidromic
activation of motor cortex (Dejean et al., 2009; Li et al., 2007).
In these studies, evoked wave in the EEG was correlated to
the positive behavioral effects. In a recent study (Gradinaru
Neuron 76, 1030–1041, D
et al., 2009), it was found that, while opto-
genetic stimulation of excitatory nerve
terminals within STN was beneficial in
improving Parkinsonian motor symp-
toms, optical inhibition or excitation
confined to STN neurons was ineffective,
raising the possibility that antidromic activation of the cortico-
STN pathway underlies the therapeutic mechanism. Our finding
that the peak antidromic frequency generated coincided with the
optimal effect of STN-DBS also supports this hypothesis. More
importantly, we showed that an antidromic spike had a strong
effect on the firing probability of the neuron immediately fol-
lowing it, and the increased mean firing rate during DBS was
primarily the effect of antidromic spikes. Our results therefore
provide the neurobiological basis of the recent findings that
highlight the importance of cortex in mediating beneficial effect
of STN-DBS. For example, by using the recorded activity to drive
the stimulation, Rosin et al. (2011) showed that short trains of
stimulation pulses were effective only if they were triggered
from cortical activity, but not from the basal ganglia. Mure
et al. (2012) showed that, in PD patients, the improved sequence
learning with STN-DBS, but not with L-3,4-dihydroxyphenylala-
nine (L-DOPA) treatment, was associated with increases in
activity in supplementary and premotor cortices. In human PD
patients, DBS of the internal globus pallidus (GPi) is also effective
in alleviating Parkinsonian symptoms (Weaver et al., 2012).
Whether a similar antidromic activation of the known cortex-
GPi projection (Naito and Kita, 1994) contributes to the thera-
peutic effect of GPi-DBS remains to be studied.
How could antidromic activation ‘‘jam’’ the synchronized
pathological bursting and beta band oscillatory rhythm in
PD? One important observation in the present study is that, as
the generation of the antidromic spike is not robust, a highly
random pattern of antidromic spikes is produced. As the firing
probability of the neuron is modified by an antidromic spike in
a biphasic manner (i.e., inhibition-excitation), the firing rate and
ecember 6, 2012 ª2012 Elsevier Inc. 1037
6-OHDA lesioned
DBS off
On
Off
Off
0
5
10
0 1 2 3 4 5 6 7 8
0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8
0 1 2 3 4 5 6 7 8 0
5
10
0
5
10
0 1 2 3 4 5 6 7 80
5
10
0 1 2 3 4 5 6 7 8 0
5
10
0 1 2 3 4 5 6 7 8
Firin
g ra
te (s
pike
s/s)
DBS on, intervals with antidromic spikes
A
DBS on, intervals without antidromic spikes
Unlesioned B
DBS on, no antidromic spikes
DBS off
Firin
g ra
te (s
pike
s/s)
Time (ms)
Time (ms) Time (ms) Time (ms)
Time (ms) Time (ms)
Time (ms) Time (ms)
a b c
a b c
On
Off
Off
Firin
g ra
te (s
pike
s/s)
Firin
g ra
te (s
pike
s/s)
0
5
0
0 1 2 3 4 5 6 7 8
DBS offffDBS offff
0 1 2 3 4 5 6 7 80
5
0
0
5
0
0 1 2 3 4 5 6 7 8
a b c
a b c
Figure 6. Antidromic Spikes Modulate
Firing Probability of CxFn
(A) Representative peristimulus rasters and
cumulative rate histograms of the firing activities of
CxFn in the 6-OHDA-lesioned sides during 2min of
125 Hz STN-DBS (‘‘on’’) and 2 min of ‘‘off’’ period
before and after DBS. Poststimulus 8 ms time
segments were aligned by setting the onset time of
stimulation to be 0ms. Those segments containing
an antidromic spike (left top panel) were separated
from those that did not contain an antidromic
spike (right top panel). In the left, during the ‘‘on’’
period, the antidromic spikes were concentrated
at around 1 ms poststimulus (time window ‘‘a’’),
followed by complete cessation of firing in time
window ‘‘b,’’ and increased firing probability in
time window ‘‘c.’’ This biphasic change following
antidromic spikes was quantified and shown in the
cumulative firing rate histogram (middle panel). In
contrast, for those time segments that did not
contain an antidromic spike (i.e., absence of spike
in ‘‘a’’), there was no change in the firing proba-
bilities in the corresponding time windows ‘‘b’’ and
‘‘c’’ (right panel). The bottom panel shows that the
firing rate of the neuron during the DBS ‘‘off’’
period was stable.
(B) In the unlesioned side, whether the DBS was
‘‘on’’ or ‘‘off,’’ no change in firing probability was
found.
Note that the lack of activity in the first ms of
the plots was due to masking by the stimulus
artifact.
See also Figures S2 and S5.
Neuron
Deep Brain Stimulation Directly Influences Cortex
rhythm of the neuron would be disrupted. We showed that each
antidromically activated CxFn was influenced by a random but
unique train of antidromic spikes that together would serve as
a powerful means to desynchronize their coherent firing.
Breaking of phase relationship among these CxFn could be
a key to this process. Although Wilson et al. (2011) proposes
that a regular stimulus pattern of DBS causes the desynchroniza-
tion, a randomly generated stimulus could also achieve the
same effect. The idea that the local circuit can be affected by
the antidromic spikes is supported by early studies that a late
response was present in cortical cells that were not antidromi-
cally activated (Phillips, 1959; Porter and Sanderson, 1964; Ste-
fanis and Jasper, 1964). There is also recent evidence from
human studies that STN-DBS has a direct effect on intracortical
neurons, modifying the balance between excitation and inhibi-
tion (Fraix et al., 2008). In fact, our data also show that anti-
dromic activation of the CxFn affected the firing of the interneu-
rons (data not shown). While our results would lend support to
the proposition that the cortex could be a therapeutic target in
PD, epidural or subdural stimulation of cortex in human beings
has been a subject of controversy. While some studies demon-
strated promising results for treating PD patients (Benvenuti
et al., 2006; Drouot et al., 2004), others were less supportive
(Kuriakose et al., 2010; Strafella et al., 2007). Similarly, the results
of transcranial magnetic stimulation were mixed (Benninger
1038 Neuron 76, 1030–1041, December 6, 2012 ª2012 Elsevier Inc.
et al., 2011; Eggers et al., 2010; Khedr et al., 2006). It is likely
that the efficacy of cortical stimulation is dependent on the
precise changes imposed on the activity of the cortical neurons,
which in turn depends on the means, locations, and parameters
of stimulation.
It should be pointed out that the observed decrease in reli-
ability of antidromic stimulation at high frequency is a nonclas-
sical observation, in contrast to the three well-accepted criteria
of antidromic spikes: fixed latency, collision, and frequency fol-
lowing (Lemon, 1984). A few factors could contribute to this
phenomenon. First, the success of antidromic invasion to the
neuronal soma in well-myelinated fibers is dependent on the
membrane voltage of the soma, as observed by Chomiak and
Hu (2007). They found that there was an overall sharp decrease
in frequency following from �40mV to �60mV within the fre-
quency range of 30–100 Hz. In the in vivo condition, it is likely
that the membrane potential of the neurons is more hyperpolar-
ized than �40mV, and therefore, one would not expect perfect
fidelity in antidromic activation. In fact, if the membrane potential
is a crucial determinant of the occurrence of antidromic spikes
in the soma, the fluctuation in the membrane potential of the
neurons could be a contributing factor to the random pattern
of antidromic spikes that we observed. Second, frequency fol-
lowing is also dependent on the degree of myelination of the
axons (Chomiak and Hu, 2007; Richardson et al., 2000). As far
Neuron
Deep Brain Stimulation Directly Influences Cortex
as we know, although the corticofugal fibers are myelinated and
fast conducting, most of the projection to the subthalamic
nucleus are minor collaterals of corticofugal fibers and are of
unmyelinated type (Afsharpour, 1985; Debanne et al., 2011).
Hence, the branch points of the collaterals could serve as low-
pass filter and increase the difficulty of antidromic invasion.
Also, as mentioned before, recruitment of inhibitory cortical
interneurons may contribute to failure of frequency following.
In conclusion, this study provided evidence that STN-DBS
antidromically activates the layer V corticofugal projection neu-
rons in the MI, which contributes to the disruption of abnormal
neural activities in the MI in PD. The unpredictable nature of anti-
dromic spikesmay hold the key to the process, a hypothesis that
needs to be verified.
EXPERIMENTAL PROCEDURES
Animals
Two groups of adult male Sprague Dawley rats weighing 250–280 gwere used,
including 17 intact and 30 hemi-Parkinsonian rats. All animal handling,
surgical, and behavior testing procedures were carried out in accordance
with university guidelines on animal ethics.
Stereotaxic Surgery
A hemi-Parkinsonian rat was generated by unilateral injection of 6-OHDA
into medial forebrain bundle (0.9% saline vehicle injection into the other
side, named as unlesioned). After two weeks’ recovery, contralateral rotation
behavior was tested for 15 min after subcutaneous injection of apomorphine
(0.5 mg/kg) and those that rotated at least 15 cycles/min were selected for
electrode implantation. Two pairs of stimulating electrodes (STABLOHM
675, CA FineWire, Grover Beach, CA) were implanted into bilateral STN (unilat-
eral in intact rats), targeting at the dorsal-lateral portion of the nucleus, which is
known to receive motor input mainly from the MI and is the site of stimulation
that generates the best motor improvement (Greenhouse et al., 2011; Roma-
nelli et al., 2004). Contralateral muscle contraction at low threshold stimulation
was indicative of the possibility that the electrodewas very near or inserted into
the internal capsule and therefore rejected for further experimentations. To
monitor the extracellular neuronal activities in the layer V of MI, two multi-
channel microwire electrode arrays, each constructed of 16 stainless steel
microwires (Plexon, Dallas, TX), were targeted at MI bilaterally (unilateral in
intact rats, ipsilateral to the stimulating electrode implantation side). The
targeted MI area corresponded to the forelimb territory, and correct loca-
tion was confirmed by epidural stimulation-induced forelimb movement. Elec-
trode placement and dopamine depletion level were confirmed histologically
postmortem.
Behavioral Assessment
The locomotor functions of both intact and hemi-Parkinsonian rats were as-
sessed by open field test two weeks after electrode implantation surgery. A
constant current isolated stimulator (Digitimer, Welwyn Garden City, Hertford-
shire, UK) delivered continuous electrical pulses to the STN electrodes at an
intensity below the threshold for induced movement (50–250 mA). The motor
performance of the hemi-Parkinsonian rat before (5 min), during (2 min), and
after (5 min) STN-DBS were compared with the spontaneous exploratory
movement (5 min) of intact rats (ANY-maze 4.70 software; Stoelting,Wood
Dale, IL). The dependence of the efficacies of DBS-STN on stimulation
frequencies (0.2, 1, 5, 10, 50, 125, 200, and 250 Hz) and pulse width (10, 20,
40, 60, 80, and 100 ms) were studied systematically. While the animals were
performing in the open field test, both extracellular spike trains and the local
field potentials (LFPs) in MI were recorded simultaneously using a 32-channel
electrophysiological data acquisition system (OmniPlex system, Plexon,
Dallas, TX).
In the behavioral assessment, muscle contractions in the contralateral face
and limb could be induced when the stimulation site was located at the lateral
N
STN border (confirmed postmortem) or the stimulation amplitude used was
high (>1 mA). Thus, contralateral muscle contraction at low threshold stimula-
tionwas indicative of the possibility that the electrodewas very near or inserted
into the internal capsule and considered unacceptable. For all other cases, the
stimulation value was set below the threshold of visible muscular contraction,
but at which it could bring behavioral improvement.
Statistics
The t tests were performed to compare the motor performance from different
groups. Paired t tests were performed on the data from hemi-Parkinsonian
rats only, comparing the STN-DBS period to both the ‘‘pre’’ and ‘‘post’’
periods. To study the dependence of behavioral improvement on stimulation
frequency and pulse width in hemi-Parkinsonian rats, an additional ANOVA
repeated-measures analysis (stimulus frequency and pulse width as
repeated-measures, respectively) followed by a LSD post hoc test was also
performed. All these behavioral test results are shown as mean ± SEM.
Electrophysiological Analysis
The stimulus artifact removal and single-unit spike-sorting process were per-
formed in the Off-line Spike Sorter V3 workspace (Plexon, Dallas, TX), using
a combination of automatic and manual sorting techniques. Burst discharge
was quantified by the Legendy surprise method. Cross-correlation analysis
was applied to study the synchronization level among CxFn. The oscillatory
rhythm in MI was measured as the spectrum of LFP using fast Fourier trans-
form at 0.2 Hz resolution. When investigating the coherence phase between
the spikes of each CxFn and the simultaneously recorded LFP, the polar histo-
gram was built by filtering the LFP into beta band (11–30 Hz).
Histology and Immunostaining
Coronal sections were cut at the STN (20 mm), MI (20 mm), SNc (20 mm), and
striatum (200 mm) by freezing microtome. Cresyl violet staining at the level of
STN and MI was performed to confirm the targets of the stimulating and
recording electrode. The sections containing the SNc and striatum were pro-
cessed by TH immunohistochemistry. The numbers of TH-positive neurons in
the SNc were counted manually, and the optical intensity of TH-immunoreac-
tivity in the striatum was quantified with Image J software.
SUPPLEMENTAL INFORMATION
Supplemental Information includes five figures and Supplemental Experi-
mental Procedures and can be found with this article online at http://dx.doi.
org/10.1016/j.neuron.2012.09.032.
ACKNOWLEDGMENTS
This work was supported by the Research Grants Council of Hong
Kong (2900336 and 478308), the NSFC/RGC Joint Research Scheme
(30931160433), and the National 973 Program (2011CB510004).
Accepted: September 17, 2012
Published: December 5, 2012
REFERENCES
Afsharpour, S. (1985). Topographical projections of the cerebral cortex to the
subthalamic nucleus. J. Comp. Neurol. 236, 14–28.
Ashby, P., Paradiso, G., Saint-Cyr, J.A., Chen, R., Lang, A.E., and Lozano,
A.M. (2001). Potentials recorded at the scalp by stimulation near the human
subthalamic nucleus. Clin. Neurophysiol. 112, 431–437.
Benabid, A.L., Pollak, P., Gross, C., Hoffmann, D., Benazzouz, A., Gao, D.M.,
Laurent, A., Gentil, M., and Perret, J. (1994). Acute and long-term effects of
subthalamic nucleus stimulation in Parkinson’s disease. Stereotact. Funct.
Neurosurg. 62, 76–84.
Benazzouz, A., Gross, C., Feger, J., Boraud, T., and Bioulac, B. (1993).
Reversal of rigidity and improvement in motor performance by subthalamic
euron 76, 1030–1041, December 6, 2012 ª2012 Elsevier Inc. 1039
Neuron
Deep Brain Stimulation Directly Influences Cortex
high-frequency stimulation in MPTP-treated monkeys. Eur. J. Neurosci. 5,
382–389.
Benninger, D.H., Berman, B.D., Houdayer, E., Pal, N., Luckenbaugh, D.A.,
Schneider, L., Miranda, S., and Hallett, M. (2011). Intermittent theta-burst
transcranial magnetic stimulation for treatment of Parkinson disease.
Neurology 76, 601–609.
Benvenuti, E., Cecchi, F., Colombini, A., and Gori, G. (2006). Extradural motor
cortex stimulation as a method to treat advanced Parkinson’s disease: new
perspectives in geriatric medicine. Aging Clin. Exp. Res. 18, 347–348.
Beurrier, C., Bioulac, B., Audin, J., and Hammond, C. (2001). High-frequency
stimulation produces a transient blockade of voltage-gated currents in subtha-
lamic neurons. J. Neurophysiol. 85, 1351–1356.
Brown, P., and Williams, D. (2005). Basal ganglia local field potential activity:
character and functional significance in the human. Clin. Neurophysiol. 116,
2510–2519.
Brown, P., and Eusebio, A. (2008). Paradoxes of functional neurosurgery: clues
from basal ganglia recordings. Mov. Disord. 23, 12–20, quiz 158.
Chang, J.Y., Shi, L.H., Luo, F., Zhang, W.M., and Woodward, D.J. (2008).
Studies of the neural mechanisms of deep brain stimulation in rodent models
of Parkinson’s disease. Neurosci. Biobehav. Rev. 32, 352–366.
Chomiak, T., and Hu, B. (2007). Axonal and somatic filtering of antidromically
evoked cortical excitation by simulated deep brain stimulation in rat brain.
J. Physiol. 579, 403–412.
Cunic, D., Roshan, L., Khan, F.I., Lozano, A.M., Lang, A.E., and Chen, R.
(2002). Effects of subthalamic nucleus stimulation on motor cortex excitability
in Parkinson’s disease. Neurology 58, 1665–1672.
Dauper, J., Peschel, T., Schrader, C., Kohlmetz, C., Joppich, G., Nager, W.,
Dengler, R., and Rollnik, J.D. (2002). Effects of subthalamic nucleus (STN)
stimulation on motor cortex excitability. Neurology 59, 700–706.
Debanne, D., Campanac, E., Bialowas, A., Carlier, E., and Alcaraz, G. (2011).
Axon physiology. Physiol. Rev. 91, 555–602.
Degos, B., Deniau, J.M., Le Cam, J., Mailly, P., and Maurice, N. (2008).
Evidence for a direct subthalamo-cortical loop circuit in the rat. Eur. J.
Neurosci. 27, 2599–2610.
Dejean, C., Hyland, B., and Arbuthnott, G. (2009). Cortical effects of subthala-
mic stimulation correlate with behavioral recovery from dopamine antagonist
induced akinesia. Cereb. Cortex 19, 1055–1063.
Deniau, J.M., Degos, B., Bosch, C., and Maurice, N. (2010). Deep brain stim-
ulation mechanisms: beyond the concept of local functional inhibition. Eur. J.
Neurosci. 32, 1080–1091.
Deuschl, G., Schade-Brittinger, C., Krack, P., Volkmann, J., Schafer, H.,
Botzel, K., Daniels, C., Deutschlander, A., Dillmann, U., Eisner, W., et al.;
German Parkinson Study Group, Neurostimulation Section. (2006). A random-
ized trial of deep-brain stimulation for Parkinson’s disease. N. Engl. J. Med.
355, 896–908.
Drouot, X., Oshino, S., Jarraya, B., Besret, L., Kishima, H., Remy, P., Dauguet,
J., Lefaucheur, J.P., Dolle, F., Conde, F., et al. (2004). Functional recovery in
a primate model of Parkinson’s disease following motor cortex stimulation.
Neuron 44, 769–778.
Eggers, C., Fink, G.R., and Nowak, D.A. (2010). Theta burst stimulation over
the primary motor cortex does not induce cortical plasticity in Parkinson’s
disease. J. Neurol. 257, 1669–1674.
Engel, A.K., and Fries, P. (2010). Beta-band oscillations—signalling the status
quo? Curr. Opin. Neurobiol. 20, 156–165.
Eusebio, A., Thevathasan, W., Doyle Gaynor, L., Pogosyan, A., Bye, E.,
Foltynie, T., Zrinzo, L., Ashkan, K., Aziz, T., and Brown, P. (2011). Deep brain
stimulation can suppress pathological synchronisation in parkinsonian
patients. J. Neurol. Neurosurg. Psychiatry 82, 569–573.
Fraix, V., Pollak, P., Vercueil, L., Benabid, A.L., and Mauguiere, F. (2008).
Effects of subthalamic nucleus stimulation on motor cortex excitability in
Parkinson’s disease. Clin. Neurophysiol. 119, 2513–2518.
1040 Neuron 76, 1030–1041, December 6, 2012 ª2012 Elsevier Inc.
Garcia, L., Audin, J., D’Alessandro, G., Bioulac, B., and Hammond, C. (2003).
Dual effect of high-frequency stimulation on subthalamic neuron activity.
J. Neurosci. 23, 8743–8751.
Goldberg, J.A., Boraud, T., Maraton, S., Haber, S.N., Vaadia, E., and Bergman,
H. (2002). Enhanced synchrony among primary motor cortex neurons in the
1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine primate model of Parkinson’s
disease. J. Neurosci. 22, 4639–4653.
Goldberg, J.A., Rokni, U., Boraud, T., Vaadia, E., and Bergman, H. (2004).
Spike synchronization in the cortex/basal-ganglia networks of Parkinsonian
primates reflects global dynamics of the local field potentials. J. Neurosci.
24, 6003–6010.
Gradinaru, V., Mogri, M., Thompson, K.R., Henderson, J.M., and Deisseroth,
K. (2009). Optical deconstruction of parkinsonian neural circuitry. Science
324, 354–359.
Greenhouse, I., Gould, S., Houser, M., Hicks, G., Gross, J., and Aron, A.R.
(2011). Stimulation at dorsal and ventral electrode contacts targeted at the
subthalamic nucleus has different effects on motor and emotion functions in
Parkinson’s disease. Neuropsychologia 49, 528–534.
Gubellini, P., Salin, P., Kerkerian-Le Goff, L., andBaunez, C. (2009). Deep brain
stimulation in neurological diseases and experimental models: from molecule
to complex behavior. Prog. Neurobiol. 89, 79–123.
Hammond, C., Bergman, H., and Brown, P. (2007). Pathological synchroniza-
tion in Parkinson’s disease: networks, models and treatments. Trends
Neurosci. 30, 357–364.
Hammond, C., Ammari, R., Bioulac, B., and Garcia, L. (2008). Latest view on
the mechanism of action of deep brain stimulation. Mov. Disord. 23, 2111–
2121.
Hashimoto, T., Elder, C.M., Okun, M.S., Patrick, S.K., and Vitek, J.L. (2003).
Stimulation of the subthalamic nucleus changes the firing pattern of pallidal
neurons. J. Neurosci. 23, 1916–1923.
Khedr, E.M., Rothwell, J.C., Shawky, O.A., Ahmed, M.A., and Hamdy, A.
(2006). Effect of daily repetitive transcranial magnetic stimulation on motor
performance in Parkinson’s disease. Mov. Disord. 21, 2201–2205.
Kita, T., and Kita, H. (2012). The subthalamic nucleus is one of multiple inner-
vation sites for long-range corticofugal axons: a single-axon tracing study in
the rat. J. Neurosci. 32, 5990–5999.
Kuriakose, R., Saha, U., Castillo, G., Udupa, K., Ni, Z., Gunraj, C., Mazzella, F.,
Hamani, C., Lang, A.E., Moro, E., et al. (2010). The nature and time course of
cortical activation following subthalamic stimulation in Parkinson’s disease.
Cereb. Cortex 20, 1926–1936.
Lee, K.H., Blaha, C.D., Harris, B.T., Cooper, S., Hitti, F.L., Leiter, J.C., Roberts,
D.W., and Kim, U. (2006). Dopamine efflux in the rat striatum evoked by elec-
trical stimulation of the subthalamic nucleus: potential mechanism of action in
Parkinson’s disease. Eur. J. Neurosci. 23, 1005–1014.
Lehmkuhle, M.J., Bhangoo, S.S., and Kipke, D.R. (2009). The electrocortico-
gram signal can be modulated with deep brain stimulation of the subthalamic
nucleus in the hemiparkinsonian rat. J. Neurophysiol. 102, 1811–1820.
Lemon, R. (1984). Methods for Neuronal Recording in Conscious Animals
(Chichester, UK: Wiley).
Li, S., Arbuthnott, G.W., Jutras, M.J., Goldberg, J.A., and Jaeger, D. (2007).
Resonant antidromic cortical circuit activation as a consequence of high-
frequency subthalamic deep-brain stimulation. J. Neurophysiol. 98, 3525–
3537.
Limousin, P., Greene, J., Pollak, P., Rothwell, J., Benabid, A.L., and
Frackowiak, R. (1997). Changes in cerebral activity pattern due to subthalamic
nucleus or internal pallidum stimulation in Parkinson’s disease. Ann. Neurol.
42, 283–291.
MacKinnon, C.D., Webb, R.M., Silberstein, P., Tisch, S., Asselman, P.,
Limousin, P., and Rothwell, J.C. (2005). Stimulation through electrodes im-
planted near the subthalamic nucleus activates projections to motor areas
of cerebral cortex in patients with Parkinson’s disease. Eur. J. Neurosci. 21,
1394–1402.
Neuron
Deep Brain Stimulation Directly Influences Cortex
Magarinos-Ascone, C., Pazo, J.H., Macadar, O., and Buno, W. (2002). High-
frequency stimulation of the subthalamic nucleus silences subthalamic neu-
rons: a possible cellular mechanism in Parkinson’s disease. Neuroscience
115, 1109–1117.
Maurice, N., Thierry, A.M., Glowinski, J., and Deniau, J.M. (2003).
Spontaneous and evoked activity of substantia nigra pars reticulata neurons
during high-frequency stimulation of the subthalamic nucleus. J. Neurosci.
23, 9929–9936.
McIntyre, C.C., and Hahn, P.J. (2010). Network perspectives on the mecha-
nisms of deep brain stimulation. Neurobiol. Dis. 38, 329–337.
Miocinovic, S., Parent, M., Butson, C.R., Hahn, P.J., Russo, G.S., Vitek, J.L.,
and McIntyre, C.C. (2006). Computational analysis of subthalamic nucleus
and lenticular fasciculus activation during therapeutic deep brain stimulation.
J. Neurophysiol. 96, 1569–1580.
Montgomery, E.B., Jr., and Gale, J.T. (2008). Mechanisms of action of deep
brain stimulation(DBS). Neurosci. Biobehav. Rev. 32, 388–407.
Mure, H., Tang, C.C., Argyelan, M., Ghilardi, M.F., Kaplitt, M.G., Dhawan, V.,
and Eidelberg, D. (2012). Improved sequence learning with subthalamic
nucleus deep brain stimulation: evidence for treatment-specific network
modulation. J. Neurosci. 32, 2804–2813.
Naito, A., and Kita, H. (1994). The cortico-pallidal projection in the rat: an
anterograde tracing study with biotinylated dextran amine. Brain Res. 653,
251–257.
Obeso, J.A., Rodriguez, M.C., Guridi, J., Alvarez, L., Alvarez, E., Macias, R.,
Juncos, J.L., and DeLong, M. (2001). Lesion of the basal ganglia and surgery
for Parkinson disease. Arch. Neurol. 58, 1165–1166.
Pasquereau, B., and Turner, R.S. (2011). Primary motor cortex of the parkinso-
nian monkey: differential effects on the spontaneous activity of pyramidal
tract-type neurons. Cereb. Cortex 21, 1362–1378.
Phillips, C.G. (1959). Actions of antidromic pyramidal volleys on single Betz
cells in the cat. Q. J. Exp. Physiol. Cogn. Med. Sci. 44, 1–25.
Porter, R., and Sanderson, J.H. (1964). Antidromic Cortical Response to
Pyramidal-Tract Stimulation in the Rat. J. Physiol. 170, 355–370.
N
Richardson, A.G., McIntyre, C.C., and Grill, W.M. (2000). Modelling the effects
of electric fields on nerve fibres: influence of the myelin sheath. Med. Biol. Eng.
Comput. 38, 438–446.
Romanelli, P., Bronte-Stewart, H., Heit, G., Schaal, D.W., and Esposito, V.
(2004). The functional organization of the sensorimotor region of the subthala-
mic nucleus. Stereotact. Funct. Neurosurg. 82, 222–229.
Rosin, B., Slovik, M., Mitelman, R., Rivlin-Etzion, M., Haber, S.N., Israel, Z.,
Vaadia, E., and Bergman, H. (2011). Closed-loop deep brain stimulation is
superior in ameliorating parkinsonism. Neuron 72, 370–384.
Sharott, A., Magill, P.J., Harnack, D., Kupsch, A., Meissner, W., and Brown, P.
(2005). Dopamine depletion increases the power and coherence of beta-oscil-
lations in the cerebral cortex and subthalamic nucleus of the awake rat. Eur. J.
Neurosci. 21, 1413–1422.
Stefanis, C., and Jasper, H. (1964). Intracellular Microelectrode Studies of
Antidromic Responses in Cortical Pyramidal Tract Neurons. J. Neurophysiol.
27, 828–854.
Strafella, A.P., Lozano, A.M., Lang, A.E., Ko, J.H., Poon, Y.Y., and Moro, E.
(2007). Subdural motor cortex stimulation in Parkinson’s disease does not
modify movement-related rCBF pattern. Mov. Disord. 22, 2113–2116.
Swann, N., Poizner, H., Houser, M., Gould, S., Greenhouse, I., Cai, W., Strunk,
J., George, J., and Aron, A.R. (2011). Deep brain stimulation of the subthalamic
nucleus alters the cortical profile of response inhibition in the beta frequency
band: a scalp EEG study in Parkinson’s disease. J. Neurosci. 31, 5721–5729.
Weaver, F.M., Follett, K.A., Stern, M., Luo, P., Harris, C.L., Hur, K., Marks,
W.J., Jr., Rothlind, J., Sagher, O., Moy, C., et al.; CSP 468 Study Group.
(2012). Randomized trial of deep brain stimulation for Parkinson disease:
thirty-six-month outcomes. Neurology 79, 55–65.
Wichmann, T., and Dostrovsky, J.O. (2011). Pathological basal ganglia activity
in movement disorders. Neuroscience 198, 232–244.
Wilson, C.J., Beverlin, B., 2nd, and Netoff, T. (2011). Chaotic desynchroniza-
tion as the therapeutic mechanism of deep brain stimulation. Front. Syst.
Neurosci. 5, 50.
euron 76, 1030–1041, December 6, 2012 ª2012 Elsevier Inc. 1041