J. Biomedical Science and Engineering, 2010, 3, 1029-1038 JBiSE
doi:10.4236/jbise.2010.310134 Published Online October 2010 (http://www.SciRP.org/journal/jbise/).
Published Online October 20 10 in SciRes. http://www.scirp.org/journal/jbise
The mode of action of electrical high frequency stimulation
Miriam Kammerer, Jonas M. Hebel, Thomas J. Feuerstein
Section of Clinical Neuropharmacology, Department of Neurosurgery, University Hospital, Freiburg, Germany.
Email: thomas.feuerstein@uniklinik-freiburg.de
Received 19 March 2010; revised 6 April; accepted 28 July 2010.
This article analyses, on the basis of the patho-
physiological grounds of various syndromes treated
with deep brain stimulation, whether there is a col-
lective explanation of the mode of action of the ap-
plied regional stimulations with high frequencies
(HFS). This proposed hypothesis assumes that HFS
selectively releases GABA. The selective GABA re-
lease can explain the efficacy and the side effects of
HFS in the various target regions according to the
maxim of the philosopher William of Ockham that
the simplest explanation is probably the correct
Keywords: HFS, DBS, Parkinson’s disease, Essential
tremor, Huntington’s disease, Depression
Deep brain stimulation (DBS) mostly reflects high fre-
quency stimulation (HFS, > 100 Hz); low frequency
stimulation (LFS, < 30 Hz) is rarely linked to the term
DBS. Parkinsonian tremor was the first syndrome which
was beneficially treated with HFS (130 Hz) in the ven-
tral intermediate thalamic nucleus, in the year 1987 [1].
Today, i.e. 22 years after this first application of HFS, its
mechanism of action is still unclear [2]. DBS is applied
in a multitude of clinical conditions, e.g. Parkinson’s
disease, Chorea Huntington, dystonia, depression, Gilles
de la Tourette syndrome, and obsessive compulsive dis-
order. For the treatment of each disorder a unique target
brain area needs to be stimulated. Therefore, many brain
target regions exist, e.g . the subthalamic nucleus (STN),
the globus pallidus medialis (GPmed), and the ventral
intermediate thalamic nucleus (VIM).
Hitherto, the following assumptions about the mode of
action of HFS and their contradictions are discussed:
HFS is thought to inactivate the stimulated structures
(see [3]). However, the decreased activity of thalamic
neurons upon GPmed-HFS [4] (evaluation in awake
monkeys) rather goes in the opposite direction. The axon
terminals from GABAergic GPmed neurons impinge on
glutamatergic thalamic neurons. Thus, their decreased
activity must be explained by a GABAA receptor- medi-
ated inhibition due to increased GABA release. The axon
terminals of GPmed neurons release GABA upon activa-
tion, not upon inhibition, by HFS. Thus, HFS may acti-
vate the neurons of the stimulated structure GPmed.
HFS activates the stimulated structures [5]. This is at
variance with the increased activity of STN neurons in
Parkinson patients [6]. In addition, HFS has been re-
ported to reduce the STN firing rate [7]. An even higher
activity due to HFS of subthalamic glutamatergic neu-
rons seems counterproductive pathophysiologically: Even
more drive of the basal ganglia output nuclei leads to
even stronger retardation of thalamic neurons, being less
active in the hypokinetic parkinsonian state anyway (see
Figure 1). Thus, HFS may inactivate the neurons of the
stimulated structure STN to alleviate hypokinesia.
Stimulation of a brain region is normally expected to re-
sult in excitatory symptoms (e.g. muscle twitchings,
flashes of light [8]). But, as Benabid et al. [2] have shown,
clinical benefits from HFS-DBS often resemble those of
earlier therapeutic lesions in the target brain areas. This
observation leads to the assumption that HFS —corre-
sponding to a functional removal of active neurons or
their effects—may be similar to an inhibition of these
neurons. As such a great variety of brain target regions,
involved neurons and treated disorders exists, it seems
quite impossible to find a single common denominator
for a possible mode of action. Nevertheless, one may ask:
Are there any mechanistics hints to solve the question of
the HFS mechanism of action?
STN, GPmed and VIM are the most often targeted DBS
regions in advanced Parkinson’s disease [9]. Frequency
is the most important parameter accounting for the thera-
peutic effects: Only HFS is efficacious, not low fre-
quency stimulation (LFS, ~20 Hz) [8]. Stimulation at
5-10 Hz even worsens Parkinsonism and no significant
M. Kammerer et al. / J. Biomedical Science and Engineering 3 (2010) 1029-1038
Copyright © 2010 SciRes. JBiSE
improvement is observed between 10 and 50 Hz [10,11].
In another brain region, the nucleus pedunculopontinus
(PPN), only LFS, not HFS, improves parkinsonian pos-
ture and gait disturbances (see 3.1). The inhibitory
GABAergic neurotransmission, including GABA neu-
rons and receptors, seems to play a predominant role in
the mechanism of action of HFS [12]. Dostrovsky et al.
[8] have proved this involvement, as local injection of
the GABAA receptor agonist muscimol into DBS target
regions in animal models imitated the corresponding
HFS effect. Consequently, HFS would affect, directly or
indirectly, GABAergic terminals, resulting in a local
release of GABA. Interestingly, only axons, which rep-
resent the most excitable components of neurons [13],
react to the widths of electrical pulses used with HFS
(60-3000 µs, see [14,15]): Chronaxies of this magnitude
are typical for nerve fibers. Compared to that, chronaxies
of cell bodies and dendrites (and also of myelin-free
synaptosomes) are approximately 10-fold higher, i.e.
1-10 ms. Therefore, HFS pulses should mainly affect
nerve fibers in areas where HFS is applied, with the
subsequent induction of neurotransmitter release from
their terminals impinging on postsynaptic cells. Their
reaction would then represent the HFS effect. Local
axon collaterals around cell bodies in an HFS target re-
gion are of course also responding to HFS if this mecha-
nism holds true. In that case, somatodendritic autore-
ceptors would respond to the transmitter released from
endings of axon collaterals. The speciality of the HFS
parameter constellation (120 to 180 Hz, 60 to 200 µs
pulse duration, current 1 mA) makes a unique mecha-
nism of action, affecting axons only, at least probable.
This is exemplified by our following recent finding using
the method of superfusion and electrical depolarization
of brain tissue. In this study we investigated, whether it
is possible to evoke [3H]-GABA and [3H]-glutamate
release from rat and human neocortical synaptosomes,
i.e. isolated nerve endings, electrically. To this end, syn-
aptosomes were pre-loaded with the triated neurotrans-
mitters and then—after incubation to take up the trans-
mitter to be investigated—superfused and stimulated. Two
different stimulation parameter constellations were ap-
plied: HFS (130 Hz, 1 mA, puls duration 0.1 ms, for 10
min) and 10 Hz, 10 mA, pulse duration 30 ms, for 1 min.
HFS did not evoke the release of [3H]-GABA (Figure 2)
or [3H]-glutamate (Figure 3) from rat neocortical syn-
aptosomes. However, the alternative parameter constel-
lation e.g. 10 Hz instead of 130 Hz, 10 mA instead of 1
mA, 30 ms instead of 0.1 ms pulse duration, application
for only 1 instead of 10 min, clearly induced the release
of both [3H]-glutamate and [3H]-GABA from synapto-
somes pre-loaded with these transmitters. Similar results
have also been found for human neocortical synapto-
somes (data not shown).
Obviously, electrical stimulations typical for HFS did
not evoke any transmitter release from neocortical syn-
aptosomes. However, another constellation of electrical
parameters, applied for only a tenth of time, clearly evoked
the synaptosomal release of [3H]-GABA or [3H]-glutamate.
This shows 1) that it is possible, as a matter of principle, to
release neurotransmitters from synaptosomes if their
higher chronaxy is translated into a much higher duration
of electrical pulses and 2) that the minimal pulse width and
electrical current together with the typical frequency of
HFS do not directly affect syn- aptosomes, i.e. nerve end-
ings. Thus, HFS may indeed excite axons exclusively; then,
transmitter release occurs not until the axonal depolariza-
tion has propagated to the nerve endings. Whether HFS is
selective for a certain neurotransmitter system, e.g. for
Figure 1. Pathophysiology of Parkinson’s disease. This figure
(adapted from [16]) illustrates the structural pathophysiologic
condition of Parkinson’s disease. Degeneration of modulating
dopaminergic neurons (dashed green line) originating from the
substantia nigra pars compacta (SNC) and projecting to the
striatum (caudate nucleus and putamen) mainly entails two
consequences: a reduced activity in formerly excited (through
dopamine D1 receptors) GABAergic interneurons (2; dashed
blue lines) and an intensified activity in formerly inhibited
(through dopamine D2 receptors) GABAergic interneurons (3;
doubled blue line). However, these two obviously oppositional
situations finally conclude in an identical effect of intensified
thalamic inhibition and a thereby increased filter function of
the thalamus. Two different pathways emanating from the
striatum and reaching the thalamus explain that. The direct
pathway (2) straightly leads from the striatum to the thalamus,
either passing the medial globus pallidus (GPmed) or the sub-
stantia nigra reticularis (SNR), whereas in the indirect pathway
(3) additionally the lateral globus pallidus (GPlat) and the glu-
tamatergic (doubled red arrows) subthalamic nucleus (STN)
are connected in series.
M. Kammerer et al. / J. Biomedical Science and Engineering 3 (2010) 1029-1038
Copyright © 2010 SciRes. JBiSE
Figure 2. [
3H]-GABA release in % of synaptosomal [3H]-
content in rat neocortex. Values in the columns represent the
number of observations. Stimulation values are given as means
with 95% confidence intervals (CI95). The significance of the
difference is indicated by asterisks: *** p < 0.001.
Figure 3. [3H]-glutamate release in % of synaptosomal [3H]-
content in rat neocortex. Values in the columns represent the
number of observation. Stimulation values are given as means
with 95% confidence intervals (CI95). The significance of the
difference is indicated by asterisks: *** p < 0.001.
GABAergic axons only, cannot be answered with these
experiments on synapto- somes.
What considerations on the basis of a selective GABA
release as HFS mechanism of action are necessary for
the different target regions? Is this unique HFS mecha-
nism of action indeed appropriate to explain consistently
and most simply why HFS acts beneficially in so many,
pathophysiologically different, syndromes? Are there
counterexamples where HFS worsens a clinical condi-
tion which would also be worsened by a selective release
of GABA? The discussion of all clinical syndromes,
which can be successfully treated with HFS without any
doubt, in the light of the proposed mechanism of action,
may illustrate the dimension of the proposed hypothesis
and possible consequences and needs in future research
in this matter.
DBS of the STN using HFS parameters improves the
cardinal symptoms of Parkinson´s disease, tremor, rigid-
ity, and bradykinesia [2]. The alleviation of the hypoki-
netic symptoms, caused by a so-called increased tha-
lamic filter function with decreased output to the neo-
cortex, can be explained comprehensively as follows.
According to Figure 1 the STN contains glutamatergic
neurons projecting to both GPmed and SNR, and further
GABAergic axon terminals originating from the GPlat. A
selective GABA release upon HFS from these fibers
impinging on glutamatergic neurons may explain the
beneficial outcome in hypokinetic patients. The released
GABA activates GABAA receptors on glutamatergic
STN neurons, resulting in a renormalization of the be-
forehand—because of a deficient GABAergic inhibi-
tion—disinhibited glutamatergic neurotransmission from
STN to GPmed and SNR (doubled red arrows). The as-
sumption of a non-selective neuronal excitation by HFS,
i.e. of GABAergic and glutamatergic fibers, would also
bring about an increased release of glutamate in the
basal ganglia output nuclei. However, this makes no
sense, as an increased glutamatergic transmission in
GPmed and SNR would finally increase the thalamic sup-
pression and therefore worsen hypokinetic symptoms.
The recently published findings of Mantovani et al.
[17] show that HFS of human neocortical slices selec-
tively induces the release of GABA, involving facilita-
tory GABAA autoreceptors may serve as an in vitro
backup for the proposed hypothesis of the mechanism of
action of DBS-HFS. Note in this context that Mantovani
et al. excluded a release of glutamate. Although the ex-
istence of a glutamate outflow due to HFS of the ventro-
lateral thalamus has been published lately [18], does this
not deductively signify an annulment of the proposed
GABA-selective action of HFS, as the reported elevation
of extracellular glutamate was not shown to reflect re-
lease from glutamatergic neurons.
Therapeutically, the most effective site for STN-HFS
is located just dorsal/dorsomedial to the STN in the area
of the pallidofugal fibers [19]. This location may contain
GABAergic fibers from the GPlat to the STN (which
should be activated by STN-HFS to induce the release of
GABA within the STN).
M. Kammerer et al. / J. Biomedical Science and Engineering 3 (2010) 1029-1038
Copyright © 2010 SciRes. JBiSE
It is possible to reduce the levodopa dosage of park-
insonian patients treated with STN-HFS by more than
50% and to accomplish an alleviation of levodopa-in-
duced hyperkinesias with this reduction [20]. Without
lowering the administered levodopa dose, STN-HFS
even worsens the dyskinesias [20-22], which can be as-
cribed to a decreased filter function of the thalamus, as
both levodopa and STN-HFS may diminish the eventual
GABAergic neurotransmission to the thalamus. Levodopa
should renormalize the patho-physiological condition in
the striatum (see Figure 1; dashed green pathways).
Concomitantly, and in accordance with our hypothesis,
STN-HFS may throttle the glutamatergic output of the
STN. Ultimately, an additionally reduced activation of
the output nuclei, GPmed and SNR, results.
3.1. LFS Treatment of Postural Imbalance and
Gait Disturbance
Postural instability and gait disturbance are two very
handicapping symptoms in Parkinson’s disease, but can
be markedly ameliorated by applying LFS to the pedun-
culopontine nucleus (PPN) [23,24]. In contrast to this,
DBS of the cholinergic and glutamatergic PPN with a
frequency of 100 Hz (approaching HFS) induced Park-
inson-like akinesia and postural imbalance in non-human
primates [25]. The described impairment may be attrib-
uted to a GABAergic inhibition of excitatory PPN neu-
rons. This assumption is reinforced by the fact that in a
monkey model of Parkinson’s disease PPN lesioning
also induced akinesia and postural instability [24]. This
lesioning presumably means a withdrawal of PPN pro-
jection fibers; this would correspond with a HFS-in-
duced local GABAergic inhibition of cholinergic and
glutamatergic PPN neurons. Further, both local applica-
tion of bicucullina GABAA receptor antagonistinto
the PPN and stimulation of the PPN with low frequency
(~20 Hz) annihilated the previous HFS-induced symp-
toms [24]. Summing up, the effects of HFS and LFS
seem to be antithetic. LFS may coincide with the ex-
pected effects of electrical stimulation of brain struc-
tures, i.e. in case of the PPN an augmentation of excita-
tory neuron activity, whereas HFS would abolish these
effects, most likely through selective GABA release and
activation of GABAA receptors on cholinergic and glu-
tamatergic PPN neurons. Altoghether, this is an example
for an in vivo correlation between HFS and GABAA re-
ceptor agonism (see [17] for an in vitro correlate). Note
that PPN-LFS primarily leads to a melioration of the
parkinsonian symptoms gait disturbance and postural
instability, but rarely of other typical parkinsonian symp-
toms like rigidity or bradykinesia [23,26]. The last-men-
tioned authors recommended a combination of bilateral
STN-HFS and PPN-LFS in appropriate Parkinson pa-
3.2. Diversity of Effects of HFS in the GPmed
The GABAergic GPmed projection neurons send their
axons to the ventral tier thalamic nuclei and to the PPN;
some of these GPmed “motor” neurons additionally pro-
ject to the CM/Pf thalamic complex [27]. The axons of
other GPmed neurons, termed “limbic” neurons by Parent
and Parent (2002), arborize principally within the lateral
habenular nucleus (LHb) with some collaterals to the
anterior thalamic nuclei. Afferents to the GPmed include
the GABAergic direct striato-pallidal monosynaptic path-
way and the glutamatergic subthalamo-pallidal part of
the indirect polysynaptic pathway. In addition, a major-
ity of GABAergic GPlat efferents have been shown to
project through the GPmed en route to the STN [27,28].
According to the literature, HFS of various GPmed tar-
gets may induce different and even contrary clinical ef-
fects. Unintentional co-stimulation of GPlat areas may
play a role here.
GPmed-HFS is reported to alleviate hypokinesias as
well as hyperkinesias [29,30]. GPmed-HFS improves ab-
normal involuntary movements, though without the pos-
sibility to reduce the dosage of levodopa essentially, in
contrast to the case of STN-HFS [31]. When the GPlat is
stimulated instead of the GPmed, HFS usually does not
improve abnormal involuntary movements [32]. When,
however, GPlat-HFS affects axons of striatal neurons of
the indirect pathway to the GPlat, HFS may increase a
too low GABAergic impulse flow in the hyperkinetic
state to improve hyperkinesias (see below). Bejjani et al.
[33] and Krack et al. [34] reported that GPmed-HFS
within the most ventral contacts, lying at the ventral
margin of, or just below, the GPmed, led to a pronounced
improvement in rigidity and a complete arrest of
levodopa-induced abnormal involuntary movements.
The anti-akinetic effect of levodopa, however, was
blocked and the patients became severely akinetic.
Stimulation of the most dorsal contacts, lying at the
dorsal border of the GPmed or inside the GPlat, usually led
to moderate improvement of off-drug akinesia and in-
duced dyskinesias in some patients. Tronnier et al. [35]
reported a reduction of dyskinesias, but a worsening of
hypokinetic parkinsonian symptoms upon GPmed-HFS, in
contrast to other studies (see [29]). Thus, multiple sites,
possibly not confined to the GPmed, but involving also
the GPlat, seem to be responsible for partly contrasting
clinical effects [33].
Obviously, the GPmed does not represent a uniform
HFS object. Two GPmed-HFS target regions have been
distinguished by Bejjani et al. [33] and by Krack et al.
[34]. There may be even more HFS targets within, and in
the close vicinity of, the GPmed:
1) HFS may affect thalamopetal axons of GPmed neu-
M. Kammerer et al. / J. Biomedical Science and Engineering 3 (2010) 1029-1038
Copyright © 2010 SciRes. JBiSE
rons and thereby improve a hyperkinetic syndrome.
2) Alternatively, HFS can reach the pallid opetal fibers
of striatal neurons of the direct pathway to increase their
too low GABAergic impulse flow in the hypokinetic
state; in this case, it alleviates hypokinesia.
3) At a still other site within the GPmed, HFS may
stimulate the en route fibers from GPlat to STN running
within the GPmed. Then, GPmed-HFS mirrors STN-HFS
and also improves hypokinesia.
4) Further, GPmed-HFS may (also) affect nearby palli-
dopetal fibers of striatal neurons of the indirect pathway
to the GPlat. In this last case, HFS increases the too low
GABAergic impulse flow in the hyperkinetic state and
improves hyperkinesia.
The pathophysiological assumptions behind (A)(D)
are the following:
a) HFS of th alamopetal axons
Abnormal involuntary movements, i.e. hyperkinesias,
of the original hypokinetic Parkinson syndrome corre-
spond with a reduced filter function of the thalamus, i.e.
an insufficient GABAergic inhibition of thalamic neu-
rons. Accordingly, a reduced neuronal activity in GPmed
during levodopa-induced dyskinesia has been shown in
parkinsonian monkeys [36]. Thus, GPmed-HFS dimin-
ishes the abnormal involuntary movements in advanced
Parkinson’s disease if the HFS-mediated selective
GABA release is paralleled by a less diminished, i.e.
normalized, GABAergic projection to the thalamus.
It was shown in human neocortex slices that HFS in-
duces action potentials in GABAergic fibers and subse-
quent terminal release of GABA with subsequent activa-
tion of facilitatory GABAA autoreceptors. GABAA re-
ceptor blockade, changing the plasmalemmal chloride
gradient of GABAA receptor channels and tetrodotoxin
(which abolishes action potentials) antagonized this
HFS-evoked GABA release [17,37]. Thus, orthodromic
action potentials may be induced in thalamopetal
GABAergic axons by HFS within the GPmed with sub-
sequent release of GABA from their thalamic terminals;
even more GABA release is due to activation by released
GABA of facilitatory GABAA autoreceptors on these
terminals. In addition, one can suppose antidromic ac-
tion potentials due to HFS. These antidromic action po-
tentials excite the soma of the GPmed neuron or travel
backwards to reach recurrent axon collaterals with sub-
sequent release of GABA in the somatodendritic region
of the GABAergic cell. GABA may increase the firing
rate of the GABAergic neuron through facilitatory
somatodendritic GABAA autoreceptors. These GABAA
autoreceptors have been demonstrated in human neocor-
tical slices [17]. Whether these somatodendritic autore-
ceptors are facilitatory, like those on GABAergic termi-
nals, or inhibitory, as usual for GABAA receptors, de-
pends on the local somatodendritic chloride gradient.
Using the pharmacological tool furosemide to change
the plasmalemmal chloride gradient, Mantovani et al.
[17] did not differentiate between the involvement of
somatodendritic and/or terminal autoreceptors in the
mode of action of HFS. In any case, the overall effect of
altering the chloride gradient was a decrease of HFS-
induced GABA release. The terminal GABAA autore-
ceptors were clearly facilitatory (as shown on isolated
nerve endings, see [17]); it may well be that the facilita-
tory terminal autoreceptors have overridden inhibitory
somatodendritic autoreceptors of minor importance for
the overall HFS-induced GABA release. Possibly, also
both terminal and somatodendritic GABAA autorecep-
tors are facilitatory and cooperate to realize the HFS-
induced GABA release. Regardless of the somatoden-
dritic autoreceptor being inhibitory or excitatory, the
facilitatory feature of the terminal GABAA autoreceptors
enabled HFS to induce an increased release of GABA. In
the case of GPmed-HFS the increase in GABA release
from terminals in the thalamus may either be the positive
net effect of facilitatory terminal and inhibitory somato-
dendritic GABAA autoreceptors or the sum of the effects
of facilitatory terminal and facilitatory somatodendritic
receptors. Boraud et al. [38] found that GPmed-HFS re-
duced the firing frequency of GPmed neurons in the
N-methyl-4-phenyl-1, 2, 3, 6-tetrahydropyridine (MPTP)-
treated parkinsonian monkey; this would correspond to
the combination of facilitatory terminal and inhibitory
somatodendritic GABAA autoreceptors.
b) HFS of pallidopetal fibers
HFS in the dorsal GPmed and/or inside the GPlat may
activate the GABAergic fibers from striatum through
GPlat to GPmed of the direct pathway (see Figure 1;
dashed blue projection from the striatum to the GPmed).
Then, the striato-pallidal GABAergic transmission of the
direct pathway is strengthened, GABAA receptors on
GPmed projection neurons are activated, i.e. the pal-
lido-thalamic neurotransmission is diminished, the filter
function of the thalamus decreases, and hypokinesia im-
proves. In the end, activating these striato-pallidal fibers
of the direct pathway should correspond to STN-HFS.
c) HFS of en route fibers from GPlat to STN
GPmed- or GPlat-HFS matches STN-HFS when axons
from GPlat neurons with terminals in the STN are stimu-
lated (see Figure 1; dashed blue projection within the
GPlat to the STN). Then, GABAergic axon terminals
within the STN will release more GABA, the glutama-
tergic subthalamo-pallidal and -nigral neurotransmis-
sions decrease, the firing rate of the basal ganglia output
nuclei is less activated, and hypokinetic parkinsonian
symptoms improve as the filter function of the thalamus
M. Kammerer et al. / J. Biomedical Science and Engineering 3 (2010) 1029-1038
Copyright © 2010 SciRes. JBiSE
d) HFS of pallidopetal fibers of striatal neurons of the
indire ct pathway
GPlat-HFS may, either intentionally or not in the
course of GPmed-HFS, target the GABAergic striato-
pallidal fibers of the indirect pathway (see Figure 1;
doubled blue projection (3) from striatum to GPlat).
In the hypokinetic parkinsonian condition these over-
active striato-pallidal fibers strongly decelerate the pal-
lido-subthalamic neurons. Then, the axons of these stri-
ato-pallidal neurons either may react to HFS with even
more increased GABA release from their terminals or
the increased GABA release is already maximal without
a further deceleration of the pallido-subthalamic neurons.
Again, also antidromic action potentials due to HFS
must be supposed; they excite the soma of the striatal
GABAergic neuron or reach recurrent axon collaterals
which subsequently release GABA in the somatoden-
dritic region in the striatum.
In hyperkinesia, the inhibition by the GABAergic
output nuclei GPmed and SNR of the thalamus is medi-
ated through the too strong dopamine D1 receptor-driven
GABAergic transmission in the monosynaptic direct
striato-pallidal and striato-nigral pathway. The resulting
reduction of the filter function of the thalamus is intensi-
fied by the D2 receptor-initiated decrease in the activity
of the polysynaptic indirect pathway to the output nuclei
with an increased GABA release in the STN and, subse-
quently, a reduced subthalamic drive of GPmed and SNR.
The proposed hypothesis predicts an increased release
of GABA upon HFS: Indeed, GPmed-HFS enhanced the
concentration of GABA in the ventricular cerebrospinal
fluid during stimulation. In addition, the GABA level
correlated with the degree of HFS-induced clinical ef-
fects against tremor, rigidity, and drug-induced dyskine-
sia [39].
3.3. HFS in the Treatment of Parkinsonian
Parkinsonian tremor can be treated by HFS in the ventral
intermediate thalamic nucleus (VIM, see 4.) and by
STN-HFS. An even better anti-tremor efficacy in Park-
inson patients seems to result from HFS in the centrum
medianum and parafascicularis thalamic nucleus (CM/Pf)
[40]. The thalamic neurons of the CM/Pf are retarded by
both GABAergic afferents from the GPmed and axon col-
laterals of GABAergic interneurons and activated by
glutamatergic afferents, e.g. from the cerebellum. Ac-
cording to these circumstances, a selective GABA re-
lease due to HFS in the CM/PF either inhibits a tremor-
transmitting cerebellar projection and/or local glutama-
tergic tremor cells.
Dyskinesias can also be improved using CM/Pf-HFS,
as shown by Krauss et al. [41], reminding of earlier
antidyskinetic outcomes of medial thalamotomies [42].
Also in this case, a local inhibition of glutamatergic
neurons due to a selective GABA release may explain
the HFS mode of action.
Excitatory afferences from the deep cerebellar nuclei
project to the VIM, which is their thalamic relay. Park-
insonian tremor [1] as well as essential tremor [43] is
improved due to the application of HFS to the VIM. Es-
sential tremor can also be improved by local injection of
the GABAA receptor agonist muscimol into the VIM, as
shown by Pahapill et al. [43]. VIM-HFS as well as the
application of muscimol results in improvement of
tremor; this leads us to presume that the selective GABA
release is the most likely HFS mode of action also in this
target area. In the VIM, a selective GABA release from
axon terminals of thalamic reticular neurons and VIM
interneurons inhibits the thalamic relay cells which are
driven by cerebellar afferents (for anatomical connec-
tions see [44]). Consequently, released GABA seems to
activate inhibitory GABAA receptors on glutamatergic
cerebellar afferents and on glutamatergic thalamic relay
A similar reasoning as for the therapy of hyperkinesias
in Parkinson’s disease, i.e. a strengthening of the
GABAergic pallido-thalamic projection, explains hypo-
thetically the efficacy of GPmed-HFS on other hyperkine-
sias, e.g. on dystonia and Huntington´s disease. One, or
even the most important, pathophysiological basis of
these hyperkinesias is also a reduced filter function of
the thalamus, to be reversed therapeutically. Besides
using HFS to treat chorea of a Huntington patient, Moro
et al. [45] also applied 40 Hz. 130 Hz improved cho-
reatic symptoms more than 40 Hz; the concomitant bra-
dykinesia, however, was rarely affected. The bradykine-
sia ameliorated with 40 Hz, admittedly at the expense of
the chorea reduction. A corresponding clinical difference
between HFS and 40 Hz-stimulation was also observed
by Fasano et al. [46]. Thus, 40 Hz induce other, possibly
opposed, effects as the HFS-typical 130 Hz.
6.1. HFS in the Treatment of Major Depression
In depression the subgenual gyrus cinguli (Brodman area
25) is metabolically overactive. This overactive metabo-
lism is being decreased due to antidepressant medication
[47]. HFS of the white matter of the subgenual gyrus
M. Kammerer et al. / J. Biomedical Science and Engineering 3 (2010) 1029-1038
Copyright © 2010 SciRes. JBiSE
cinguli was applied to reduce this elevated activity and
successfully improved treatment-resistant major depres-
sion. Therefore white matter of the subgenual gyrus
cinguli-HFS may lead to an increased release of GABA
from axon terminals in the Brodman area 25. GABA
then activates inhibitory GABAA receptors on overactive
postsynaptic neurons which calms these neurons.
6.2. Suicidality as Side-effect of STN-HFS
A depression-like behaviour is aggravated in the forced
swim test due to STN-HFS; the forced swim test is a
widely used and validated rodent model of depression.
On the level of neuronal activities, the firing rate of
5-HT neurons in the dorsal raphe nucleus (NDR) of rats
is inhibited following STN-HFS [48]. Muscimol, being
infused into the STN, imitated the effects of STN-HFS
on the firing rate of 5-HT- neurons. Voon et al. [49] have
recently confirmed that suicide is one of the most im-
portant risks for mortality following STN-HFS in ad-
vanced Parkinson’s disease. This serious adverse effect’s
pathophysiology may allow us to draw conclusions
about the mode of action of HFS.
Anatomically, the following connections between
STN and the 5-HT neurons of the NDR exist (see
[50,51]; : excitation, : inhibition, ncl. habenulae lat-
eralis: LHb, GABA interneurons of NDR: NDRGABA,
5-HT neurons of NRD: NDR5-HT):
The physiological condition is reflected by the fol-
lowing chain of neuronal impacts:
The condition in patients suffering from Parkinson’s
disease, however is different. In the hypokinetic state the
STN is disinhibited, which changes the above-mentioned
( means increased, () decreased, excitation;
means increased, () decreased, inhibition):
STN  GPmed LHb () NDRGABA () NDR5-HT.
A reduced inhibition of NDR 5-HT neurons is the re-
sult of the disinhibited STN.
The condition after STN-HFS, however, may vary as
Obviously, STN-HFS increases the inhibition of 5-HT
neurons of the NDR which results in a lowering of the
serotonergic neurotransmission to cortical areas. Defi-
ciencies in the monoamine neurotransmission is the cur-
rent hypothesis underlying major depression. Conse-
quently, this decrease may explain the increased sui-
cidality of Parkinson patients after STN-HFS.
Not only 5-HT neurons in the NDR, but also nora-
drenergic neurons in the locus coeruleus (LC), which
project to the cortical areas, are influenced by STN-HFS.
Pathophysiological condition in the hypokinetic Park-
inson syndrome:
STN  GPmed LHb () LCGABA () LCNA.
Condition in Parkinson patients after STN-HFS:
An increased inhibition due to a lowered noradrener-
gic neurotransmission also promotes the occurrence of
depression [52].
Note that the coincidence of a decrease of the sero-
tonergic as well as the noradrenergic neurotransmission
may lead to a substantially increased risk of the occur-
rence of depression.
GPmed-HFS is also used in the treatment of Parkin-
son’s disease. Does suicidality also occur in GPmed-HFS
as an adverse effect? Rodriguez-Oroz et al. [53] have
compared the clinical occurence of depression in Park-
inson patients treated with these two different HFS
methods, i.e. STN-HFS and GPmed-HFS. A lower rate of
depressions after GPmed-HFS compared to STN-HFS
was found. This could be the result of a decreased inhi-
bition of NDR5-HT after GPmed-HFS which ends in an
undiminished serotonergic neurotransmission to cortical
areas. Therefore, depression is not likely to occur after
Condition of Parkinson patients after GPmed-HFS:
According to Servello et al. [54] it is possible to suc-
cessfully treat patients suffering from Tourette syndrome
by applying HFS to the centrum medianum/ parafas-
cicular nucleus (CM/Pf) of the thalamus and to the ven-
tral oral anterior thalamic nucleus (Voa). Although the
pathophysiological knowledge about the exact network
of neurotransmitters acting in the Gilles de la Tourette
syndrome is limited, one may assume a GABAergic in-
hibition of glutamatergic (CM/Pf and Voa) and choliner-
gic (CM/Pf) neurons through GABAA receptors. The
GABAergic afferents in this case come from the GPmed.
While applying HFS to the CM/Pf and Voa, a terminal
release of GABA is induced in afferent fibers to these
nuclei and, by this, neurons in the HFS target structures
are inhibited. Thus, the above stated assumption leads to
the proposal that CM/Pf- and Voa-efferents, projecting to
both striatum and neocortex, there may trigger the Gilles
de la Tourette syndrome.
Corresponding to a recent publication by Greenberg et al.
[55] DBS of the ventral internal capsule (VC) and the
ventral striatum (VS) with HFS parameters meliorates
the symptoms of patients with Obsessive Compulsive
Disorder. Also here the particular pathophysiology is
M. Kammerer et al. / J. Biomedical Science and Engineering 3 (2010) 1029-1038
Copyright © 2010 SciRes. JBiSE
unclear, but, in compliance with our hypothesis of the
mode of functioning of HFS, GABA release in the VS as
well as from VC fiber endings would take place. Now,
on the one hand, axon collaterals from GABAergic inter
or projection neurons of the VS could be excited by HFS
and, on the other hand, postsynaptic neurons in target
areas of VC fibers could be influenced in an inhibitory
manner by the released GABA.
Various human epilepsies have also been experimentally
treated with DBS and by subdural neocortical stimula-
tion (target regions: hippocampus, cerebellum, thalamus,
STN, neocortex; see [56]). Comparing the efficacies, a
higher rate of electrical stimulation approaches in animal
models of epilepsies than of the corresponding clinical
applications have displayed anticonvulsant properties
(e.g. STN-HFS against absence-like seizures, cortex
piriformis-LFS in kindled animals, hippocampus-HFS
and -LFS). Taken together, electrical stimulation meth-
ods in the treatment of epilepsies seem to be more re-
mote from a common clinical application than the other
clinical syndromes mentioned above. Mechanistically,
however, just epilepsies could offer interesting aspects
for the implementation of HFS inducing selective GABA
release, as regards the pathophysiological role of the
opponents GABA and glutamate in these disorders.
Various treatment locations and options for HFS against
different neurological and psychiatric syndromes are
discussed above; there are detailed pathophysiological
conceptions for most of these syndromes and, addition-
ally for the suicidality following STN-HFS [49]. The
hypothesis of a selective GABA release due to HFS is in
line with these conceptions (see underlined pathophysi-
ological basis in the following) by explaining the effica-
cies and side effects of HFS according to the stimulated
regions. (A) Augmented filter function of the thalamus
(see 3.): GABA, released in the STN from axon termi-
nals of neurons from the GPlat, reduces the disinhibition
of the glutamatergic neurotransmission from STN to
GPmed and to SNR and thus (re-) normalizes their tha-
lamopetal projections. (B) Reduced filter function of the
thalamus (see 3.2, 5.): GPmed-HFS increases the GABA-
ergic transmission of projection neurons to the thalamus.
(C) In contrast to STN-HFS, GPmed-HFS is not depres-
siogenic, according to the pathophysiological conception
of the transmission from GPmed to LHb to both NDR and
LC. (D) In opposition to PPN-LFS (3.1), PPN-HFS im-
pairs gait disturbance and postural instability of the Park-
inson syndrome, as HFS induces a GABAergic inhibi-
tion of excitatory PPN neurons. (E) CM/Pf-HFS and
VIM-HFS are effective against tremor and dyskinesias
(3.3, 4.) by a GABAergic inhibition of glutamatergic
thalamic neurons. (F) HFS in the subgenual gyrus cin-
guli inhibits through GABAergic axon terminals overac-
tive neurons of the Brodmann area 25 in depression.
The error probability of a correct explanation of the
overall mechanism of action of HFS (selective GABA
release) with regard to its clinical effects may be as-
sessed as follows: Together, six independent arguments
have been listed (A-F). If this independence of the six
arguments is accepted, then a single probability of only
39.5% has to be assumed: The validity of the hypothesis
“a selective GABA release explains argument (A) or (B)
or … or (F)” corresponds to a significant collective ex-
planation for the mode of action of HFS. The error
probability for this collective explanation is in that case
p = 0.049 = (1 – 0.395)6. Thus, we can state that a selec-
tive GABA release significantly explains the mode of
action of HFS. The proposed hypothesis corresponds to
an optimal simplicity in explaining the observed clinical
effects since the collective explanation is in any case
simpler than another one which assumes different modes
of actions of HFS in different target regions.
We should add that the individual evidence of the se-
lective GABA release following HFS in the various con-
ditions should be separately demonstrated in spite of the
present consideration of a collective explanation.
Disclosure/Conflict-of-Interest Statement: This re-
search was conducted in the absence of any commercial
or financial relationships that could be construed as a
potential conflict of interest.
[1] Benabid, A.L., Pollak, P., Louveau, A., Henry, S. and
Rougemont, J.D. (1987) Combined (thalamotomy and
stimulation) stereotactic surgery of the VIM thalamic nu-
cleus for bilateral Parkinson disease. Applied Neurophysi-
ology, 50(1-6), 344-346.
[2] Benabid, A.L., Chabarde, S., Mitrofanis, J. and Pollak, P.
(2009) Deep brain stimulation of the subthalamic nucleus
for the treatment of Parkinson disease. Lancet Neurology,
8(1), 67-81.
[3] Beurrier, C., Bioulac, B., Audin, J. and Hammond, C.
(2001) High-frequency stimulation produces a transient
blockade of voltage-gated currents in subthalamic neurons.
Journal of Neurophysiology, 85(4), 1351-1356.
[4] Anderson, M.E., Postupna, N. and Ruffo, M. (2003) Ef-
fects of high-frequency stimulation in the internal globus
pallidus on the activity of thalamic neurons in the awake
monkey. Journal of Neurophysiology, 89(2), 1150-1160.
[5] Montgomery, E.B. and Baker, K.B. (2000) Mechanisms of
deep brain stimulation and future technical developments.
M. Kammerer et al. / J. Biomedical Science and Engineering 3 (2010) 1029-1038
Copyright © 2010 SciRes. JBiSE
Neurological Research, 22(3), 259-266.
[6] Levy, R., Ashby, P., Hutchison, W.D., Lang, A.E., Lozano,
A.M. and Dostrovsky, J.O. (2002) Dependence of subtha-
lamic nucleus oscillations on movement and dopamine in
Parkinson’s disease. Brain, 125, 1196-1209.
[7] Meissner, W., Leblois, A., Hansel, D., Bioulac, B., Gross,
C.E., Benazzouz, A. and Boraud T. (2005) Subthalamic
high frequency stimulation resets subthalamic firing and
reduces abnormal oscillations. Brain, 128, 2372-2382.
[8] Dostrovsky, J.O. and Lozano, A.M. (2002) Mechanisms
of deep brain stimulation. Movement Disorders, 17(3)
[9] Garcia, L., Allesandro, G.D., Bioulac, B. and Hammond,
C. (2005) High-frequency stimulation in Parkinson’s dis-
ease: More or less? Trends in Neuroscience, 28(4),
[10] Moro, E., Esselink, R.J., Xie, J., Hommel, M., Benabid,
A.L. and Pollak, P. (2002) The impact on Parkinson’s
disease of electrical parameter settings in STN stimulation.
Neurology, 59(5), 706-713.
[11] Rizzone, M., Lamotte, M., Bergamasco, B., Tavella, A.,
Torre, E., Faccani, G., Melcarne, A. and Lopiano, L. (2001)
Deep brain stimulation if the subthalamic nucleus in
Parkinson’s disease: Effects of variation in stimulation
parameters. Journal of Neurology, Neurosurgery, and
Psychiatry, 71(2), 215-219.
[12] Benabid, A.L., Benazzous, A. and Pollak, P. (2002)
Mechanisms of deep brain stimulation. Movement Disor-
ders, 17(3), S73-S74.
[13] Nowak, L.G. and Bullier, J. (1998) Axons, but not cell
bodies, are activated by electrical stimulation in cortical
grey matter. I. evidence from chronaxie measurements.
Experimental Brain Research, 118(4), 477-488.
[14] Holsheimer, J., Demeulemeester, H., Nuttin, B. and Sutter,
P.D. (2000) Identification of the target neuronal elements
in electrical deep brain stimulation. European Journal of
Neuroscience, 12(12), 4573-4577.
[15] Holsheimer, J., Dijkstra, E.A., Demeulemeester, H. and
Nuttin, B. (2000) Chronaxie calculated from current-duration
and voltage-duration data. Journal of Neuroscience Methods,
97(1), 45-50.
[16] Feuerstein, T.J. (2009) Antiparkinsonmittel. Pharmako-
therapie des morbus Parkinson. In: Aktories, K., Förster-
mann, U., Hofmann, F., Starke, K., Eds., Allgemeine und
Spezielle Pharmakologie and Toxikologie, Urban & Fischer,
München, 299-305.
[17] Mantovani, M., Moser, A., Haas, C.A., Zentner, J. and
Feuerstein, T.J. (2009) GABAA autoreceptors enhance
GABA release from human neocortex: Towards a me-
chanism for high-frequency stimulation (HFS) in brain?
Naunyn-Schmiedeberg’s Archives of Pharmacology, 380
(1), 45-58.
[18] Chang, S., Shon, Y.M., Agnesi, F. and Lee, K.H. (2009)
Microthalamotomy effect during deep brain stimulation:
Potential involvement of adenosine and glutamate efflux.
Conference Proceedings: 31st Annual International Con-
ference of the IEEE EMBS, 3294-3297.
[19] Plaha, P., Shlomo, Y.B., Patel, N.K. and Gill, S.S. (2006)
Stimulation of the caudal zona incerta is superior to
stimulation of the subthalamic nucleus in improving con-
tralateral Parkinsonism. Brain, 129, 1732-1747.
[20] Moro, E., Scerrati, M., Romito, L.M., Roselli, R., Tonali,
P. and Albanese, A. (1999) Chronic subthalamic nucleus
stimulation reduces medication requirements in Parkin-
son’s disease. Neurology , 53(1), 85-91.
[21] Krack, P., Pollak, P., Limousin, P., Hoffman, D., Xie, J.,
Benazzouz, A. and Benabid, A.L. (1998) Subthalamic nu-
cleus or internal pallidal stimulation in young onset Park-
inson’s disease. Brain, 121, pp. 451-457.
[22] Limousin, P., Pollak, P., Hoffmann, D., Benazzouz, A.,
Perret, J.E. and Benabid, A.L. (1996) Abnormal involun-
tary movements induced by subthalamic nucleus stimula-
tion in parkinsonian patients. Movement Disorders, 11 (3 ) ,
[23] Plaha, P. and Gill, S.S. (2005) Bilateral deep brain stimu-
lation of the pedunculopontine nucleus for Parkinson’s
disease. Neuroreport, 16(17), 1883-1887.
[24] Pereira, E.A., Muthusamy, K.A., Pennington, N.D., Joint,
C.A. and Aziz, T.Z. (2008) Deep brain stimulation of the
pedunculopontine nucleus in Parkinson’s disease. Pre-
liminary experience at Oxford. British Journal of Neuro-
surgery, 22(1), S41-S44.
[25] Nandi, D., Liu, X., Winter, J.L., Aziz, T.Z. and Stein, J.F.
(2002) Deep brain stimulation of the pedunculopontine
region in the normal non-human primate. Journal of
Clinical Neuroscience, 9(2), 170-174.
[26] Stefani, A., Lozano, A.M., Peppe, A., Stanzione, P., Galati,
S., Tropepi, D., Pierantozzi, M., Brusa, L., Scarnati, E. and
Mazzone, P. (2007) Bilateral deep brain stimulation of the
pedunculopontine and subthalamic nuclei in severe Park-
inson’s disease. Brain, 130, 1596-1607.
[27] Parent, M. and Parent, A. (2002) Axonal collateralization
in primate basal ganglia and related thalamic nuclei.
Thalamus & Related systems, 2(1), 71-86.
[28] Sato, F., Lavallée, P., Lévesque, M. and Parent, A. (2000)
Single-axon tracing study of neurons of the external seg-
ment of the globus pallidus in primate. Journal of Com-
parative Neurology, 417(1), 17-31.
[29] Volkmann, J. (2004) Deep brain stimulation for the treat-
ment of Parkinson’s disease. Journal of Clinical Neuro-
physiology, 21(1), 6-17.
[30] Siegfried, J. and Lippitz, B. (1994) Bilateral chronic elec-
trostimulation of ventroposterolateral pallidum: A new
therapeutic approach for alleviating all parkinsonian symp-
toms. Neurosurg ery, 35(6), 1126-1129.
[31] Perlmutter, J.S. and Mink, J.W. (2006) Deep brain stimu-
lation. Annual Review of Neuroscience, 29, 229-257.
[32] Peppe, A., Pierantozzi, M., Altibrandi, M.G., Giacomini,
P., Stefani, A., Bassi, A., Mazzone, P., Bernardi G. and
Stanzione, P. (2001) Bilateral GPi DBS is useful to reduce
abnormal involuntary movements in advanced Parkin-
son’s disease patients, but its action is related to modality
and site of stimulation. European Journal of Neurology, 8
(6), 579-586.
[33] Bejjani, B., Damier, P., Arnulf, I., Bonnet, A.M., Vidailhet,
M., Dormont, D., Pidoux, B., Cornu, P., Marsault C. and
Agid, Y. (1997) Pallidal stimulation for Parkinson’s dis-
ease. Two targets? Neurology, 49(6), 1564-9156.
[34] Krack, P., Pollak, P., Limousin, P., Hoffman, D., Benaz-
zouz A. and. Bas, J.F.L., Koudsie, A. and Benabid, A.L.
(1998) Opposite motor effects of pallidal stimulation in
Parkinson’s disease. Annals of Neurology, 43(2), 180-192.
[35] Tronnier, V.M., Fogel, W., Kronenbuerger, M. and Stein-
vorth, S. (1997) Pallidal stimulation: An alternative to pal-
M. Kammerer et al. / J. Biomedical Science and Engineering 3 (2010) 1029-1038
Copyright © 2010 SciRes. JBiSE
lidotomy? Journal of Neurosurgery, 87(5), 700-705.
[36] Papa, S.M., Desimone, R., Fiorani, M. and Oldfield, E.H.
(1999) Internal global pallidus discharge is nearly sup-
pressed during levodopa-induced dyskinesias. Annals of
Neurology, 46(5), 732-738.
[37] Mantovani, M., Velthoven, V.V., Fuellgraf, H., Feuerstein,
T.J. and Moser, A. (2006) Neuronal electrical high fre-
quency stimulation enhances GABA outflow from human
neocortical slices. Neurochemistry International, 49(4),
[38] Boraud, T., Bezard, E., Bioulac, B. and Gross, C. (1996)
High frequency stimulation of the internal Globus pallidus
(GPi) simultaneously improves parkinsonian symptoms
and reduces the firing frequency of GPi neurons in the
MPTP-treated monkey. Neuroscience Letters, 210(1),
[39] Ogura, M., Nakao, N., Nakai, E., Uematsu, Y. and Itakura,
T. (2004) The mechanism and effect of chronic electrical
stimulation of the globus pallidus for treatment of Park-
inson disease. Journal of Neurosurgery, 100(6), 997-1001.
[40] Peppe, A., Gasbarra, A., Stefani, A., Chiavalon, C., Pier-
antozzi, M., Fermi, E., Stanzione, P., Caltagirone, C. and
Mazzone, P. (2008) Deep brain stimulation of CM/PF of
thalamus could be the new elective target for tremor in
advanced Parkinson’s Disease? Parkinsonism & Related
Disorders, 14(6), 501-504.
[41] Krauss, J.K., Pohle, T., Weigel, R. and Burgunder, J.M.
(2002) Deep brain stimulation of the centre median-
parafascicular complex in patients with movement disor-
ders. Journal of Neurology, Neurosurgery and Psychiatry,
72(4), 546-548.
[42] Ohye, C. and Shibazak, T. (2001) Lesioning the thalamus
for dyskinesia. Stereotactic and Functional Neurosurgery,
77(1-4), 33-39.
[43] Pahapill, P.A., Levy, R., Dostrovsky, J.O., Davis, K.D.,
Rezai, A.R., Tasker, R.R. and Lozano, A.M. (1999)
Tremor arrest with thalamic microinjections of muscimol
in patients with essential tremor. Annals of Neurology,
46(2), 249-252.
[44] Strafella, A., Ashby, P., Munz, M., Dostrovsky, J.O.,
Lozano, A.M. and Lang, A.E. (1997) Inhibition of volun-
tary activity by thalamic stimulation in humans: Rele-
vance for the control of tremor. Movement Disorders,
12(5), 727-737.
[45] Moro, E., Lang, A.E., Strafella, A.P., Poon, Y.Y., Arango,
P., Dagher, A., Hutchison, W.D. and Lozano, A.M. (2004)
Bilateral globus pallidus stimulation for Huntington’s
disease. Annals of Neurology, 56(2), 290-294.
[46] Fasano, A., Mazzone, P., Piano, C., Quaranta, D., Soleti F.
and Bentivoglio, A.R. (2008) GPi-DBS in Huntington’s
disease: Results on motor function and cognition in a
72-year-old case. Movement Disorders, 23(9), 1289-1292.
[47] Mayberg, H.S., Lozano, A.M., Voon, V., McNeely, H.E.,
Seminowicz, D., Hamani, C., Schwalb, J.M. and. Ken-
nedy, S.H. (2005) Deep brain stimulation for treat-
ment-resistant depression. Neuron, 45(5), 651-660.
[48] Temel, Y., Boothman, L.J., Blokland, A., Magill, P.J.,
Steinbusch, H.W., Vandewalle, V.V. and Sharp, T. (2007)
Inhibition of 5-HT neuron activity and induction of de-
pressive-like behavior by high-frequency stimulation of
the subthalamic nucleus. Proceedings of the National
Academy of Sciences of the United States of America,
104(43), 17087-17092.
[49] Voon, V., Krack, P., Lang, A.E., Lozano, A.M., Dujardin,
K., Schüpbach, M., Ambrosia, J.D., Thobois, S., Tamma,
F., Herzog, J., Speelman, J.D., Samanta, J., Kubu, C.,
Rossignol, H., Poon, Y.Y., Cyr, J.A.S., Ardouin, C. and
Moro, E. (2008) A multicentre study on suicide outcomes
following subthalamic stimulation for Parkinson’s disease.
Brain, 131, 2720-2728.
[50] Ferraro, G., Montalbano, M.E., Sardo, P. and Grutta, V.L.,
(1996) Lateral habenular influence on dorsal raphe neu-
rons. Brain Research Bulletin, 41(1), 47-52.
[51] Smith, Y., Bevan, M.D., Shink, E. and Bolam, J.P. (1998)
Microcircuitry of the direct and indirect pathways of the
basal ganglia. Neuroscience, 86(2), 353-387.
[52] Sartorius, A. and Henn, F.A. (2007) Deep brain stimula-
tion of the lateral habenula in treatment resistent major
depression. Medical Hypotheses, 69(6), 1305-1308.
[53] Oroz, M.C.R., Obeso, J.A., Lang, A.E., Houeto, J.L., Pol-
lak, P., Rehncrona, S., Kulisevsky, J., Albanese, A., Volk-
mann, J., Hariz, M.I., Quinn, N.P., Speelman, J.D., Guridi,
J., Zamarbide, I., Gironell, A., Molet, J., Sedano, B. P.,
Pidoux, B., Bonnet, A.M., Agid, Y., Xie, J., Benabid, A.L.,
Lozano, A.M., Cyr, J.S., Romito, L., Contarino, M.F.,
Scerrati, M., Fraix, V. and Blercom, N.V. (2005) Bilateral
deep brain stimulation in Parkinson’s disease: A multicen-
tre study with 4 years follow-up. Brain, 128, 2240-2249.
[54] Servello, D., Porta, M., Sassi, M., Brambilla, A., Robert-
son, M.M., (2008) Deep brain stimulation in 18 patients
with severe Gilles de la Tourette syndrome refractory to
treatment: The surgery and stimulation. Journal of Neu-
rology, Neurosurgery and Psychiatry, 79(2), 136-142.
[55] Greenberg, B.D., Gabriels, L.A., Jr, D.A.M., Rezai, A.R.,
Friehs, G.M., Okun, M.S., Shapira, N.A., Foote, K.D.,
Cosyns, P.R., Kubu, C.S., Malloy, P.F., Salloway, S.P.,
Giftakis, J.E., Rise, M.T., Machado, A.G., Baker, K.B.,
Stypulkowski, P.H., Goodman, W.K., Rasmussen, S.A.
and Nuttin, B.J. (2010) Deep brain stimulation of the ven-
tral internal capsule/ventral striatum for obsessive-compulsive
disorder: worldwide experience. Molecular Psychiatry,
15(1), 64-79.
[56] Gubellini, P., Salin, P., Goff, L.K.L. and Baunez, C. (2009)
Deep brain stimulation in neurological diseases and ex-
perimental models: From molecule to complex behaviour.
Progress in Neurobiology, 89(1), 79-123.