Stimulation of the mesencephalic locomotor region for gait recovery after stroke: DBS for Gait Recovery

Abstract

Objective:
One-third of all stroke survivors are unable to walk, even after intensive physiotherapy. Thus, other concepts to restore walking are needed. Since electrical stimulation of the mesencephalic locomotor region (MLR) is known to elicit gait movements, this area might be a promising target for restorative neurostimulation in stroke patients with gait disability. The present study aims to delineate the effect of high-frequency stimulation of the MLR (MLR-HFS) on gait impairment in a rodent stroke model.

Methods:
Male Wistar rats underwent photothrombotic stroke of the right sensorimotor cortex and chronic implantation of a stimulating electrode into the right MLR. Gait was assessed using clinical scoring of the beam walking test and videokinematic analysis (CatWalk™) at baseline and on days 3 and 4 after experimental stroke with and without MLR-HFS.

Results:
Kinematic analysis revealed significant changes in several dynamic and static gait parameters resulting in overall reduced gait velocity. All rats exhibited major coordination deficits during the beam walking challenge and were unable to cross the beam. Simultaneous to the onset of MLR-HFS, a significantly higher walking speed and improvements in several dynamic gait parameters were detected by the Catwalk™-system. Rats regained the ability to cross the beam unassisted showing a reduced number of paw slips and misses.

Interpretation:
MLR-HFS can improve disordered locomotor function in a rodent stroke model. It may act by shielding brainstem and spinal locomotor centers from abnormal cortical input after stroke, thus allowing for compensatory and independent action of these circuits. This article is protected by copyright. All rights reserved.

Full text

RESEARCH ARTICLE
Stimulation of the Mesencephalic
Locomotor Region for Gait
Recovery After Stroke
Felix Fluri, MD,
1
Uwe Malzahn, PhD,
2
Gy€
orgy A. Homola, PhD,
3
Michael K. Schuhmann, PhD,
1
Christoph Kleinschnitz, MD,
1
* and
Jens Volkmann, MD, PhD
1
Objective: One-third of all stroke survivors are unable to walk, even after intensive physiotherapy. Thus, other con-
cepts to restore walking are needed. Because electrical stimulation of the mesencephalic locomotor region (MLR) is
known to elicit gait movements, this area might be a promising target for restorative neurostimulation in stroke
patients with gait disability. The present study aims to delineate the effect of high-frequency stimulation of the MLR
(MLR-HFS) on gait impairment in a rodent stroke model.
Methods: Male Wistar rats underwent photothrombotic stroke of the right sensorimotor cortex and chronic implanta-
tion of a stimulating electrode into the right MLR. Gait was assessed using clinical scoring of the beam-walking test
and video-kinematic analysis (CatWalk) at baseline and on days 3 and 4 after experimental stroke with and without
MLR-HFS.
Results: Kinematic analysis revealed significant changes in several dynamic and static gait parameters resulting in
overall reduced gait velocity. All rats exhibited major coordination deficits during the beam-walking challenge and
were unable to cross the beam. Simultaneous to the onset of MLR-HFS, a significantly higher walking speed and
improvements in several dynamic gait parameters were detected by the CatWalk system. Rats regained the ability to
cross the beam unassisted, showing a reduced number of paw slips and misses.
Interpretation: MLR-HFS can improve disordered locomotor function in a rodent stroke model. It may act by shield-
ing brainstem and spinal locomotor centers from abnormal cortical input after stroke, thus allowing for compensatory
and independent action of these circuits.
ANN NEUROL 2017;00:000–000
Motor deficits are the most common symptoms after
ischemic stroke. Approximately one-third of stroke
survivors suffer from impaired mobility and cannot walk
1 year after an ischemic cerebrovascular event.
1
Gait after
stroke is characterized by a slow and asymmetrical walk-
ing pattern, with reduced stride length and a prolonged
swing phase of the affected limb.
2
There are different
rehabilitation strategies after stroke, such as physical ther-
apy,
3
neurodevelopmental training, and motor relearning
programs.
4
However, the overall effect of physiotherapy
in chronic stroke survivors is modest,
5
and there are no
pharmacological or interventional alternatives. This is
even more remarkable as locomotion relies on phyloge-
netically well-conserved brainstem and spinal circuits,
which themselves remain unaffected by a hemispheric
lesion. These mesencephalic and spinal central pattern
generators (CPGs) produce the alternating rhythmic neu-
ral signals necessary for stepping behavior
6
and the anti-
gravitational postural adjustments required for walking.
7
In animals, CPGs are capable of sustaining coordinated
View this article online at wileyonlinelibrary.com. DOI: 10.1002/ana.25086
Received Sep 5, 2016, and in revised form Dec 29, 2016, Jun 20, 2017, and Oct 19, 2017. Accepted for publication Oct 20, 2017.
Address correspondence to Dr Fluri, Department of Neurology, University Hospital W€
urzburg, Josef-Schneider Strasse 11, 97080 W€
urzburg, Germany.
E-mail: felix.fluri@gmx.ch
From the
1
Department of Neurology, University Hospital W€
urzburg;
2
Institute of Clinical Epidemiology and Biometry, University of W€
urzburg; and
3
Department of Diagnostic and Interventional Neuroradiology, University Hospital W €
urzburg, W€
urzburg, Germany.
*Current address for Dr Kleinschnitz: Department of Neurology, University Hospital Essen, Essen, Germany.
Additional supporting information can be found in the online version of this article.
V
C2017 American Neurological Association 1

gait even after complete deafferentation from cortical
input.
8,9
Hence, the interesting question arises whether
gait impairment after stroke could reflect a dysfunc-
tional descending input to brainstem and spinal central
pattern generators,
10
rather than a loss of function of
these pathways. This would open the possibility of neu-
romodulatory therapy such as deep brain stimulation
(DBS), which is highly effective in treating dysfunc-
tional motor circuit activity in Parkinson disease
11
or
dystonia.
12
Here, we explore the restorative potential of DBS
of the mesencephalic locomotor region (MLR) in a rat
model of hemiplegic stroke. We induced gait deficits in
Wistar rats by inflicting a photothrombotic lesion to the
right sensorimotor cortex,
13
and evaluated stimulation
effects of the MLR by video-kinematic gait analysis and
behavioral assessment during a gait challenge. Photo-
thrombosis is a widely used stroke model in rodents with
the advantage of lesioning a defined brain region, thus
allowing probing the response of the disabled brain site
and interconnections to a targeted intervention. The
MLR was chosen as a target for neuromodulation
because it is involved in the initiation and control of
gait.
14
A recently published study suggests that the MLR
of rats consists of noncholinergic, predominantly gluta-
matergic cells localized in the lateral pontine tegmentum,
which form a strip extending from the ventrolateral peri-
aqueductal gray matter to the region ventromedial to the
pedunculopontine tegmental nucleus (PTg) with ipsilat-
eral predominace.
15
Electrical or pharmacological stimu-
lation of this region elicits locomotor behavior even in
decerebrated animals such as cats,
8
rats,
16
salamanders,
17
and nonhuman primates.
18
Moreover, there is clinical
experience with MLR stimulation in patients with Par-
kinson disease,
19–21
in principle making it feasible to
translate results from rodents to humans.
Materials and Methods
Animals
All experiments were performed in adult (250–275g, 6–8 weeks
old) male Wistar rats (n 520; Charles River, Sulzfeld,
Germany). Rats were acclimatized for 1 week in an animal facil-
ity and housed in a room with con trolled temperature (22 6
0.5 8C) under a 12/12-hour light/dark cycle. They were allowed
free access to food and water. All animal experiments were
approved by the institutional review board of Julius Maximilian
University, W€urzburg and by the local authorities of lower
Frankonia (Regierung von Unterfranken, W€urzburg, Germany).
Induction of Photothrombotic Stroke
Animals were divided into 2 groups: the first underwent photo-
thrombotic stroke only (n 510), whereas the second was sub-
jected to both photothrombosis and electrode insertion into the
MLR (n 510, see below), to exclude any behavioral effect of
electrode implantation into the MLR.
Rats were anesthetized with 2.5% isoflurane during the
surgery. Body temperature was maintained at 37 60.5 8Cbya
feedback-controlled heating system. A photothrombotic cerebral
stroke was induced in all rats as follows.
22
A template with an
aperture (10 35mm) for the light source was put on the
exposed skull 5mm anterior to 5mm posterior and 0.5mm to
5.5mm lateral to bregma, an area which corresponds to the sen-
sorimotor cortex (Fig 1).
23
A light guide was placed over the
aperture. Rose Bengal (0.5ml; Sigma, St Louis, MO) in NaCl
0.9% (10mg/ml) was administered intravenously, and the brain
was illuminated (KL1500LCD; Olympus, Tokyo, Japan)
through the intact skull for 15 minutes. Immediately after this
procedure, a microelectrode was implanted in half of the ani-
mals (see section below).
Microelectrode Implantation
For high-frequency stimulation of the MLR (MLR-HFS),
monopolar microelectrodes (catalogue # UE-PSEGSECN1M;
FHC, Bowdoin, ME) were used in this study. To avoid electro-
chemical neurotoxicity, electrodes made of platinum/iridium
with a mean impedance of 0.93MX(range 50.8–1.2MX) were
used. Electrodes were implanted close to the dorsal part of the
MLR ipsilateral to the lesion (coordinates: 7.8mm posterior,
2.0mm lateral, and 5.8mm ventral to the bregma) as described
in detail elsewhere.
24
By using the aforementioned coordinates,
the tip of the electrode was placed slightly above the dorsal part
of the MLR (Fig 2A), which avoids the destruction of this
small structure but still ensures an effect of stimulation on the
MLR. The electrode was implanted ipsilateral to the photo-
thrombotic stroke for the following reason. In a recent study by
Bachmann et al,
25
unilateral injection of FastBlue into the left
rostral medulla oblongata resulted in predominant retrograde
labeling of the left MLR and left motor cortical areas and to a
lesser extent of the contralateral cortex, indicating a largely
uncrossed organization of the corticomesencephalic–spinal loco-
motor circuit (as corroborated by Matsumura et al
26
in
macaque monkey).
Five stainless steel screws (M1.6; length, 3mm; Hummer
& Rieß, Nuremberg, Germany) were inserted in boreholes
without penetrating the dura overlaying the brain surface. A
custom-made plug (GT-Labortechnik, Arnstein, Germany) was
put on the pin of the electrode, and the ground wire of the
plug was connected with one of the screws. To fix the elec-
trode/plug with the bone screws, dental cement was applied on
the skull and molded around the electrode/plug by forming a
small cap. Wound edges were closed with a suture at the front
and behind the cap. Thereafter, animals were allowed to wake
up.
Behavioral Testing
Rats were trained for 7 days to traverse a horizontal wooden
beam (90cm long, 9mm wide, 70cm above ground). At the
same time, they learned to cross the runway of the CatWalk
system (Noldus, Wageningen, the Netherlands), a video-based
ANNALS of Neurology
2 Volume 00, No. 0

analysis system to assess static and dynamic gait parameters (for
a complete description of this method, see Hamers et al
27
). On
the last day of training, traversing the wooden beam and cross-
ing the CatWalk system (3 runs per animal) were recorded;
these measurements were used as baseline values. Induction of
photothrombotic stroke and implantation of the electrode were
performed 1 day later (ie, on day 8 after the beginning of train-
ing). Three days after intervention, locomotor behavior was first
investigated without stimulation using the CatWalk system.
Three hours later, the same experiment was carried out with
FIGURE 1: Visualization of the photothrombotic stroke and the electrode tip. (A) Representative coronal T2-weighted (T2w)
magnetic resonance scans revealing the photothrombotic stroke (hyperintense area) in the right sensorimotor cortex of a rat
brain. (B) Macroscopic view of a rat brain after removal from the skull. The black arrows indicate the photothrombotically
induced lesion. Primary motor cortex (M1), blue framed; secondary motor cortex (M2), green framed; primary somatosensory
cortex, forelimb, yellow framed; primary somatosensory cortex, hindlimb, brown framed (according to Paxinos and Watson’s
rat brain atlas
23
). (C) Brain sections in 3 planes (coronal [top]; sagittal [middle]; horizontal [bottom]) of an averaged brain gener-
ated from T2w scans of the rats used in this study. The overlapping size and site of the photothrombotic lesion is displayed by
a heatmap on the right hemisphere. The sections are superimposed on the corresponding atlas template.
23
The heatmap color
red represents a low overlapping of 10%, whereas yellow indicates an overlapping of 100%. Within the M1, overlapping of
the lesions is almost 100%. Cg1 5cingulate cortex, area 1; Cg2 5cingulate cortex, area 2; FrA 5frontal association cortex;
MPtA 5medial parietal association cortex; PrL5prelimbic cortex; S1BF 5primary somatosensory cortex, barrel field;
S1DZ 5primary somatosensory cortex, dysgranular zone; S1FL 5primary somatosensory cortex, forelimb region; S1HL 5pri-
mary somatosensory cortex, hindlimb region; S1J 5primary somatosensory cortex, jaw region; S1Tr 5primary somatosensory
cortex, trunk region; S1ULp 5primary somatosensory cortex, upper lip region; V2ML 5secondary visual cortex, mediolateral.
Fluri et al: DBS for Gait Recovery
Month 2017 3

HFS (frequency 5130Hz, pulse length 560 microseconds,
pulse shape 5monophasic square wave pulses) using the stimu-
lus generator STG 4002 (Multichannel Systems, Reutlingen,
Germany). This device includes a large voltage compliance
range of 120V, as well as constant current stimulation as a par-
ticular feature. In the current mode, the device is able to adjust
the voltage to changes in tissue impedance, and thus provides a
constant current output at the electrode. Before starting gait
analyses, a threshold current intensity was determined for each
animal by observing spontaneous locomotor behavior as
described recently.
25
The current threshold for stimulation-
evoked locomotion was determined by beginning at 20 mA and
then increasing the intensity in 10 mA steps until the maximal
locomotion was seen. Switching on and off the device tested
reproducibility of the stimulus-induced locomotion. Thereafter,
the lowest current-evoking locomotion was chosen for further
testing. In the present study, the lowest current intensity result-
ing in increased locomotor activity was 40 mA in all tested ani-
mals. Three crossings of the CatWalk runway without
interruption/hesitation were required for a valid kinematic gait
analysis in each animal. Data were analyzed using the CatWalk
XT 10 software.
The 5 most widely used gait parameters in recently pub-
lished studies on locomotion after stroke
28,29
were analyzed,
namely step cycle, swing speed, and duty cycle (ie, dynamic
paw parameters) as well as stride length and contact area (ie,
static paw parameters, Supplementary Table 1).
The beam-walking task was performed 4 days after pho-
tothrombosis using the same parameters for HFS as during
CatWalk testing. Three traverses per animal were performed
and video was recorded. The time passing between the first and
the last step on the beam was taken to calculate the gait speed.
Fine motor coordination and balance were further determined
using a 7-point nonparametric scale and animals were scored as
follows
30
:15unable to traverse or falls off the beam;
25unable to traverse the beam but able to maintain balance
on the beam; 3 5able to traverse the beam by dragging the
affected limb; 4 5able to traverse the beam and—at least
once—to place the affected limb on the horizontal surface of
the beam; 5 5the affected limbs are used in <50% of its steps
FIGURE 2: Verification of electrode placements. Consecutive brain sections encompassing the electrode site were used for
fluorescent in situ hybridization to visualize choline acetyl-transferase (ChAT)
1
neurons of the pedunculopontine tegmental
area (PTg) and c-Fos
1
neurons indicating the stimulation site. Both ChAT
1
and c-Fos
1
cell groups as well as the electrode site
(dots) were mapped onto atlas drawings of the rat brain. The relationship of electrode sites to cholinergic neurons of the PTg
(A) and to the c-Fos
1
neurons (B) are visualized by cloud diagrams. It is of note that ChAT
1
neurons rarely expressed c-Fos in
our study. The numbers below the drawings indicate the anterior–posterior distance to the bregma. The light gray areas indi-
cate the PTg, the dark gray areas the cuneiform nucleus (Cn).
ANNALS of Neurology
4 Volume 00, No. 0

on the beam; 6 5able to traverse the beam by using the
affected limbs (contralateral to the lesion and implanted elec-
trode) for >50% of its steps along the beam; 7 5able to tra-
verse the beam normally with no more than 2 foot slips. To
examine whether the use of the left fore- and hindlimb changes
during MLR-HFS after photothrombotic stroke, paw slips and
misses off the beam (1 point per fault) were counted before
photothrombotic stroke and thereafter under HFS conditions.
Both the CatWalk analysis and the beam-walking test
were also carried out in rats subjected to photothrombotic
stroke alone, to investigate whether electrode implantation into
the MLR influences locomotor behavior.
Measurement of Lesion Volume
Lesion size was visualized using T2-weighted (T2w) magnetic
resonance imaging (MRI) on a 3.0T scanner (MAGNETOM
Trio; Siemens, Erlangen Germany). T2w scans were acquired
with turbo spin-echo sequences (echo time 5105 milliseconds,
repetition time 52,100 milliseconds) and infarct volume was
determined using ImageJ Analysis Software 1.45s (National
Institutes of Health, Bethesda, MD; http://rsb.info.nih.gov/ij/);
the hyperintense lesion on each scan (1mm thick) was traced
manually and the areas were then summed and multiplied by
the slice thickness.
To compare the location and size of the photothrombotic
lesion among all animals, an average brain of these animals was
generated and all lesions were overlapped in a color-coded heat-
map on this brain as follows. T2w images were brain-extracted
with the brain extraction tool of FSL (FMRIB, Oxford, UK)
optimized for rodent brains and additionally corrected manu-
ally. All lesions were segmented manually in original data. Rat
brains were registered with FLIRT (FMRIB). The lesion masks
were transformed according to the individual brain registra-
tions. The sum of all lesions was overlaid on the normalized
average brain data using a color lookup table. Thereafter, 3
brain sections representing each plane were superimposed on
the corresponding atlas template.
23
Immunohistochemistry
After deep anesthesia, rats were killed by decapitation and the
brains were harvested rapidly and immediately frozen at
220 8C. Coronal sections (12 mm thick) were cut using a cryo-
stat (Leica 3050; Leica Microsystems, Wetzlar, Germany). Sec-
tions encompassing the MLR were stained with hematoxylin
and eosin to visualize the anatomic locations of the electrode
tip. The localization of the stimulation sites was assessed by
choosing 2 approaches. First, the relationship of the electrode
tip location was mapped out with respect to the cholinergic
neurons of the PTg; second, c-Fos expression sites were com-
pared to the localization of cholinergic neurons.
To identify choline acetyl-transferase (ChAT)-positive
neurons of the PTg and to visualize the expression of c-Fos,
fluorescent in situ hybridization of sections encompassing the
MLR and lesion due to the electrode tip were performed using
the RNAscope Multiplex Fluorescent v2 Assay according to the
manufacturer’s instructions (Advanced Cell Diagnostics, Milan,
Italy; catalogue # 323100). Target probes for c-Fos (RNAscope
probe Rn-Fos, catalogue # 403591) and ChAT (RNAscope
probe Rn-Chat-C2, catalogue # 430111-C2) were designed by
Advanced Cell Diagnostics. After amplification and label appli-
cation, sections were counterstained with 4,6-diamidino-2-phe-
nylindole (Sigma-Aldrich, St Louis, MO; catalogue # D9542).
Images were acquired with a Leica MDi8 microscope (magnifi-
cation 5403). Finally, ChAT
1
cells of the PTg as well as c-
Fos
1
cells around the stimulation site of each animal were
delineated as a cloud onto atlas drawings of consecutive (corre-
sponding) brain sections. These cell groups were then related to
the distal end of the electrode trajectory.
Statistical Analysis
For gait speed and number of step cycles measured by the Cat-
Walk system, individual averages of each rat were calculated
over 3 runs for each time point (ie, measurements before and
after photothrombotic stroke, as well as during MLR-HFS) and
used to get group means and standard deviations (SDs). Gait
speed and number of step cycles were further analyzed using
repeated measures analysis of variance (ANOVA) with Green-
house–Geisser corrections as appropriate for sphericity viola-
tions. Post hoc analyses were performed with Tukey multiple
comparison test.
Additional gait parameters (ie, step cycle, swing speed,
and duty cycle as well as stride length and contact area) mea-
sured by the CatWalk system were analyzed as raw values in
relation to instantaneous body velocity.
31
This was necessary,
because most gait parameters change as a function of speed
32
and photothrombotic stroke reduces gait velocity, such that the
intervention itself would act as a confounder. In a first step, we
plotted scattergraphs of each parameter against body velocity
and compared the mostly nonlinear distributions under the 3
different treatment conditions visually. We then conducted
global and velocity-restricted group comparisons (repeated mea-
sures ANOVA and post hoc ttests for paired samples with
Bonferroni correction) at slow (16–30cm/s), medium (30–
65cm/s), and fast (65–150cm/s) body speed.
For the beam-walking test, statistical differences of gait
velocity and scores before photothrombosis and 4 days after
intervention under MLR-HFS were calculated using the 2-
tailed paired ttest. All values are presented as mean 6SD with
95% confidence intervals (CIs). Probability values <0.05 were
considered to indicate statistical significance. Statistical Package
for the Social Sciences (SPSS 17.0; IBM, Armonk, NY) soft-
ware was used for statistical analysis.
Results
Baseline Characteristics
One animal died in each group during the experiment
and thus had to be excluded from the analyses. T2w
scans revealed a photothrombotic lesion in all animals
encompassing the right sensorimotor cortex (see Fig 1A,
B). On T2w scans, lesion size did not differ significantly
between both groups (mean lesion volume: first group,
Fluri et al: DBS for Gait Recovery
Month 2017 5

72.8 66.4mm
3
vs second group, 84.9 69.6mm
3
;
p50.32). To further determine size and site of the pho-
tothrombotic stroke, all scanned brains were coregistered
and a template was calculated; thereafter, the degree of
overlapping of the photothrombotic lesions was visual-
ized by a color-coded heatmap on the template. Whereas
the primary motor cortex (M1) was affected to almost
100% in all animals, the degree of lesional overlapping
was decreased within the secondary motor cortex (M2)
and the primary somatosensory cortex representing the
hindlimb (estimated 75%) and was even less within the
primary somatosensory cortex representing the forelimb
(estimated 50%; see Fig 1C).
Immediately after intervention, as well as before
kinematic gait testing (ie, 3 days after the intervention),
all animals exhibited normal cage mobility and no coor-
dination deficits were observed in the use of the affected
left forepaw during food uptake.
Hematoxylin and eosin staining revealed some vari-
ability of electrode placement (see Fig 2A). In 6 animals,
the electrodes were placed close to or within the cunei-
form nucleus (Cn; ie, 27.80mm from the bregma),
whereas 2 animals exhibited a deviation of the tip posi-
tion into the anterior direction 120 mm from the Cn (ie,
27.68mm from bregma). The electrode tip of a third
animal was found at the lower right outer border of the
Cn, 27.92mm from bregma. It was decided to keep all
animals in the analysis using an intention-to-treat
approach, because the optimal mesencephalic site of
stimulation was not yet defined.
Histological Analysis of the Neurostimulation
Effects
Fluorescent in situ hybridization of consecutive brain sec-
tions were performed to visualize the lesion due to the
electrode tip, the ChAT
1
neurons of the PTg, and c-
Fos
1
neurons around the stimulation site of each animal.
Then, brain sections were mapped onto atlas drawings of
the rat brain to show the relationship of electrode sites to
cholinergic neurons of the PTg (see Fig 2A) as well as to
FIGURE 3: Assessment of locomotor behavior before and
after photothrombotic stroke without and during high-
frequency stimulation of the mesencephalic locomotor
region (MLR-HFS; day 4 after intervention). (A) Whereas no
locomotion was seen after photothrombotic stroke (PT)
without MLR-HFS, gait velocity changed after stroke almost
to the baseline values when animals were stimulated in the
MLR-HFS. *p<0.001 (95% confidence interval [CI] 520.31–
24.66), #p<0.001 (95% CI 5222.82 to 216.48); ns 5not
significant (p>0.05, 95% CI 520.72 to 6.40); error bars
indicate standard deviation; 2-tailed paired ttest. (B) Beam-
walking score, assessed according to a 7-point scale (see
Materials and Methods section). After photothrombotic
stroke, MLR-HFS restored gait coordination and balance sig-
nificantly compared to the test condition without MLR-HFS.
*p<0.001 (95% CI 55.25–6.36); #p<0.001 (95% CI 526.38
to 23.49); ns 5not significant (p>0.05, 95% CI 520.33 to
22.06); error bars indicate standard deviation; 2-tailed
paired ttest. (C) Effect of MLR-HFS on affected fore- and
hindpaw. Whereas there was no difference between fore-
paw and hindpaw regarding faults before photothrombotic
stroke, rats made significantly more faults with the hindpaw
than with the forepaw after photothrombotic stroke even
during MLR-HFS. *p50.037 (95% CI 522.54 to 20.13); ns,
not significant (p>0.05, 95% CI 520.29 to 0.29); error bars
indicate standard deviation; 2-tailed paired ttest. FL 5fore-
limb; HL 5hindlimb; non-stim 5assessment after PT, no
MLR-HFS; stim 5assessment after PT, with MLR-HFS.
ANNALS of Neurology
6 Volume 00, No. 0

the c-Fos
1
neurons (see Fig 2B). The more rostrally
localized electrodes (ie, 27.68mm anterior–posterior to
the bregma) had less effect on gait speed and explain
why velocities of 25.4cm/s and 25.6cm/s were mea-
sured in these 2 animals. Different elements such as
cell somata (ie, glutamatergic as well as cholinergic
neurons) as well as axons and dendrites might be acti-
vated by applying HFS.
Beam-Walking Test
The behavioral outcome after stroke with and without
MLR-HFS was evaluated using the beam-walking test.
This test allows evaluation of coordination and
integration of paw movements after skilled gait training.
On day 7 of training (ie, before photothrombotic stroke),
the average speed of the beam traversing was 22.5 6
0.8cm/s (Supplementary Video 1). Four days after photo-
thrombotic stroke, all animals demonstrated paw
coordination deficits and were unable to traverse the
beam without assistance (Supplementary Video 2). When
MLR-HFS was applied, coordinated locomotion was
restored instantaneously and an average speed of 19.7 6
1.2cm/s for unassisted beam traversing was recorded
(Supplementary Video 3, Fig 3A), which was similar to
the gait velocity measured before photothrombotic stroke
(p50.096).
TABLE 1. Comparison of Gait Parameters Using the CatWalk System: Changes in Locomotor Variables Mea-
sured Before and After Photothrombotic Stroke
LH LF RH RF
Parameter Overall
Mean
Constrained
Velocities
a
Overall
Mean
Constrained
Velocities
a
Overall
Mean
Constrained
Velocities
a
Overall
Mean
Constrained
Velocities
a
Body speed ## ns ## ns ## " ## ns
ns ns ## ##
ns ns ns ns
Step cycle, s "ns ns ns ns ns ns "
"" ns "" "
ns ns ns ns
Duty cycle ## # ns ns #ns ns ns
## ns ## #
## ns ns ns
Stride
length, cm
#ns ns ns ## ns ## ns
ns ns ns ns
ns ## ns ##
Swing speed,
cm/s
## ns ## ## ## ns ## ns
## ## ## ##
## ns ns ns
Contact
area, cm
2
## ## ## ## ## ns ## ns
## ## ## ##
## ## ns ##
A significant increase/decrease of a gait parameter is indicated by "/#(p<0.05) and ""/## (p<0.01). Individual numeric values for each parame-
ter and paw are outlined in Supplementary Table 2.
a
Slow (top), medium (middle), and fast (bottom).
LF 5left forelimb; LH 5left hindlimb; ns 5not significant; RF 5right forelimb; RH 5right hindlimb.
Fluri et al: DBS for Gait Recovery
Month 2017 7

Changes in locomotor skills and balance were fur-
ther determined using a 7-point scale. On day 7 of train-
ing, mean score was 7; only 1 animal slipped with the
forepaw and another with the hindpaw when crossing
the beam. Four days after photothrombotic stroke, skilled
walking on the beam was first tested without HFS in all
animals. Only 1 animal was able to maintain balance on
the beam; all others fell off the beam. After applying
MLR-HFS, skilled locomotion improved significantly in
all animals; one of them returned even to a score of 7
(mean), whereas the animal with least effect of MLR-
HFS regained a mean score of 4.6 (see Fig 3B).
To determine whether MLR-HFS exerts a more pow-
erful effect on the fore- or hindlimb, we assessed paw slips
and misses off the beam before photothrombotic stroke
and thereafter under MLR-HFS conditions. The number
of slips measured for forepaw and hindpaw did not differ
before intervention. Whereas no locomotion was visible
and thus this parameter was not evaluable after photo-
thrombotic stroke without MLR-HFS, significantly fewer
paw slips and misses off the beam were observed for the
forepaw compared to the hindpaw during MLR-HFS (see
Fig 3C).
No difference in locomotor behavior was found
between animals with photothrombotic stroke alone and
those subjected to both photothrombosis and electrode
implantation (ie, both groups were no longer able to tra-
verse the beam on day 4 after the intervention), exclud-
ing a clinically relevant impact of MLR microlesioning
by electrode implantation.
TABLE 2. Comparison of Gait Parameters Using the CatWalk System: After Photothrombosis without and
with Mesencephalic Locomotor Region High-Frequency Stimulation
LH LF RH RF
Parameter Overall
Mean
Constrained
Velocities
a
Overall
Mean
Constrained
Velocities
a
Overall
Mean
Constrained
Velocities
a
Overall
Mean
Constrained
Velocities
a
Body speed "" ns "" ns "" ns "" ns
ns "" ns ""
""""""
Step cycle, s ## ns ## ns #ns ns ns
#ns ns ns
## ns #ns
Duty cycle ## ns ns ns ns ns #ns
ns ns ns ns
ns ns ns ns
Stride length, cm "ns ns ns "" ns "" ns
ns ns "" ns
"" "" ns ns
Swing speed, cm/s "" ns "" # "" # "" ns
ns ns ns ns
"" ns ns ns
Contact area, cm
2
ns ## # ## ns ## ns ##
ns ns ns ns
ns ns ns ns
A significant increase/decrease of a gait parameter is indicated by "/#(p<0.05) and ""/## (p<0.01). Individual numeric values for each parame-
ter and paw are outlined in Supplementary Table 2.
a
Slow (top), medium (middle), and fast (bottom).
LF 5left forelimb; LH 5left hindlimb; ns 5not significant; RF 5right forelimb; RH 5right hindlimb.
ANNALS of Neurology
8 Volume 00, No. 0

CatWalk Analyses
Locomotor impairments of fore- and hindpaws after stroke
and MLR-HFS–related changes of gait were quantified
using the CatWalk system. A similar total number of step
cycles before and after photothrombotic stroke was mea-
sured (3.9 60.4 vs 3.9 60.3; p50.86, 95% CI 520.43
to 0.51). After induction of stroke, MLR-HFS did not
change significantly the number of step cycles compared to
the nonstimulated state (3.9 60.3 vs 3.7 60.6; p50.49,
95% CI 520.40 to 0.77). Gait velocity (mean) was 43.26
6.6cm/s in “healthy” animals, which was significantly
reduced after photothrombosis (31.7 69.0cm/s; p50.007,
95% CI 54.11–18.9). When MLR-HFS was applied after
photothrombotic stroke, a significant increase in gait veloc-
ity was observed (43.8 612.6cm/s vs 31.7 69.0cm/s;
p50.04, 95% CI 5223.5 to 20.74).
When comparing gait parameters before and after
photothrombosis (without HFS), mean step cycle of the
right paws and the left forepaw did not change signifi-
cantly (Table 1), whereas mean stride length, swing
speed, and contact area of all paws (except for the stride
length of the left forepaw) were significantly decreased.
With respect to stride length, the right paws exhibited
the largest deficits. The velocity constrained analysis
revealed that these deficits were largely observed at
medium gait velocity, but not at slow velocity.
TABLE 3. Comparison of Gait Parameters Using the CatWalk System: Before Photothrombosis and Thereafter,
When Mesencephalic Locomotor Region High-Frequency Stimulation Was Applied
LH LF RH RF
Parameter Overall
Mean
Constrained
Velocities
a
Overall
Mean
Constrained
Velocities
a
Overall
Mean
Constrained
Velocities
a
Overall
Mean
Constrained
Velocities
a
Body speed ns ns ns ns ns "" ns ns
ns ns ##
"" " ns ""
Step cycle, s ns ns ns ns ns ns ns ns
ns ns "" ns
## # ns ns
Duty cycle ## ns ns ###ns ns ns
## ns ## ##
## ns ns ns
Stride
length, cm
ns ns ns ns ns ns ns ns
ns ns ns ns
"" ns ns ns
Swing
speed, cm/s
## ns ns ## ## ns ns ns
## ## ## ##
#ns ns ns
Contact
area, cm
2
## ## ## ## ## ## ## ##
## ## ## ##
## ns ns #
A significant increase/decrease of a gait parameter is indicated by "/#(p<0.05) and ""/## (p<0.01). Individual numeric values for each parame-
ter and paw are outlined in Supplementary Table 2.
a
Slow (top), medium (middle), and fast (bottom).
LF 5left forelimb; LH 5left hindlimb; ns 5not significant; RF 5right forelimb; RH 5right hindlimb.
Fluri et al: DBS for Gait Recovery
Month 2017 9

When MLR-HFS was applied after stroke, the over-
all mean value of step cycle decreased significantly for all
paws except for the right forepaw, whereas mean values
of swing speed increased significantly (Table 2). Mean
values for stride length increased significantly for the
right paws, whereas the value of the left hindpaw
remained unchanged. Mean values of the contact area—a
static paw parameter—did not change significantly dur-
ing MLR-HFS.
Next, we compared gait parameters before photo-
thrombosis and thereafter when MLR-HFS was applied
(Table 3). There was no significant difference with
respect to step cycle and stride length between the base-
line and stimulated stroke condition and duty cycle and
swing speed of the forepaws. In contrast, contact area of
all 4 paws, as well as duty cycle and swing speed of the
hindpaws, remained significantly reduced after photo-
thrombotic stroke despite MLR-HFS.
Discussion
In the present study, we examined the effect of MLR-
HFS on stroke-related locomotor deficits in rats, which
underwent photothrombotic lesioning of the right senso-
rimotor cortex and implantation of a stimulation elec-
trode into the MLR ipsilateral to the infarction. We
verified that the MLR is a site of action of HFS by dem-
onstrating immunohistochemically that glutamatergic as
well as cholinergic cells of the dorsal part of the MLR
but also cells in adjacent areas expressed c-Fos after
MLR-HFS. It is difficult, however, to assign the stimula-
tion effect to particular nuclei or neural elements within
this region, because monopolar stimulation with a cur-
rent intensity of 40 mA, as used in this experiment, may
excite neural elements (ie, myelinated axons) within a
radius of 500 to 700 mm from the electrode tip.
33
Which
of these elements alone or in combinations contribute to
the observed behavioral responses can only be answered
in future investigations using cell-type–specific stimula-
tion techniques (eg, opto- or pharmacogenetics).
Gait analysis on day 3 after photothrombosis
revealed impairment of dynamic gait parameters caused
by paresis and—to a lesser extent—by coordination defi-
cits of the contralateral fore- and hindlimb during video-
kinematic assessment. The deficits were subtle and barely
visible during spontaneous cage locomotion. Only chal-
lenging tests such as the beam-walking task, requiring
nonparetic paws and unimpaired interlimb coordination
for maintaining body balance on a narrow path, revealed
a clinically relevant locomotor deficit in cortically
lesioned animals. Implantation of the stimulating elec-
trode alone had no impact on poststroke gait symptoms,
whereas acute HFS of the MLR through the chronically
implanted electrode resulted in an immediate restoration
of the ability to cross the test beam without assistance.
The most prominent behavioral changes induced
by MLR-HFS were an increased gait velocity (127.6%)
as revealed by kinematic analysis as well as a significant
amelioration of the skilled walking on the beam. Interest-
ingly, when investigating the left fore- and hindpaw
regarding slips off the wooden beam, a significantly
higher recovery of the forepaw compared to the hindpaw
was seen when cortically lesioned animals traversed the
beam under MLR-HFS. This might be explained by less
lesioning of the forelimb than hindlimb representation
within sensorimotor cortex as shown by the heatmaps.
However, the differences of lesion size and site between
the somatosensory hind- and forelimb representation
were small, and cortical representations vary widely
among individual rats. Alternatively, one might argue
that the microelectrode was implanted in a section of the
MLR representing the forepaw. Again, this seems
unlikely, because previous studies suggest a nonsomato-
topic and rather mixed body representation in the
MLR.
26
We would therefore like to forward the follow-
ing alternative hypothesis. Descending projections of the
MLR target the medullary and pontine reticular forma-
tion.
34
Recently, Esposito and coworkers have shown that
a distinct brainstem nucleus in the ventral part of the
medullary reticular formation plays a crucial role in con-
trolling motor activity of forepaws by demonstrating that
this brainstem area is connected to a subset of forelimb-
innervating spinal motor neurons.
35
Furthermore, their
experiments showed that a larger number of brainstem
nuclei are connected with forelimb than with hindlimb
motor neurons,
35
which might explain—in part—the
better response of the left forepaw to MLR-HFS.
Although the MLR has no direct axonal projections to
the somatomotor cortex, it is nevertheless indirectly con-
nected to the cortex via a relay in the thalamus,
36
which
might contribute to the modulation of the motor func-
tion of the forelimb.
During MLR-HFS, gait velocity increased signifi-
cantly in cortically lesioned animals. This finding might
be of clinical relevance, because independent community
ambulation of stroke patients has been shown to require
a certain degree of gait velocity (ie, 0.80m/s).
37
Addition-
ally, an increase of gait velocity augments the degree of
ambulatory activity, which is low in stroke survivors.
38
However, gait velocity induced by MLR-HFS varied
greatly among animals, as indicated by the large SD. It is
of note that stimulation parameters were always the same
in each animal, such that we suspect electrode placement
in relation to the individual MLR as an important con-
tributing factor. The anatomical extent of the MLR
ANNALS of Neurology
10 Volume 00, No. 0

overlaps with a region ventromedial to the PTg,
15
the
mesencephalic reticular nucleus, PTg, and Cn in
rodents.
39
Most of the electrodes were placed in the cen-
ter or toward the ventral margin of the Cn. Electrical
stimulation of the ventral margin of the Cn has been
shown to evoke alternating hindlimb movements in pre-
collicular–postmamillary decerebrated cats or a change in
locomotor behavior from fast walking to gallop in pre-
collicular–premamillary decerebrated cats.
40
In the pre-
sent study, 2 of the electrodes were detected at the far
rostral pole of both the Cn/PTg and the MLR. The cur-
rent–distance relation
33
may result in less intense stimu-
lation of the MLR cell populations controlling
locomotor speed (especially the population in the region
ventromedial to the PTg) with such an electrode loca-
tion, which—in turn—would explain a more modest
increase of speed after HFS. Additionally, repetitive elec-
trical stimulation of the dorsal part of the PTg in decere-
brated cats elicited stepping movements of the hindlimb;
however, these repetitive stimuli subsequently attenuated
locomotion along with a decrease in muscle tone.
40
Alto-
gether, the variability of placement of electrodes in the
present study indicates that many sites in the midbrain
may have some effect on locomotion, but not necessarily
the same, which is in line with the study published by
Takakusaki et al.
40
The presented findings are also ham-
pered by the small sample size of 9 animals.
Altogether, this study demonstrates that MLR-HFS
can ameliorate gait disability in a rat model of hemiple-
gic stroke and that a unilateral stimulation of the MLR
(ie, ipsilateral to the photothrombotic stroke) is sufficient
to improve quadrupedal walking. This emphasizes the
restorative potential of mesencephalic and spinal motor
circuits supporting locomotion, which may be unlocked
by neuromodulation therapy. We propose that MLR-
HFS shields the mesencephalic and downstream locomo-
tor systems from aberrant cortical input after stroke, and
allows for autonomous function of these circuits. The
nature and origin of the dysfunctional input activity
remains enigmatic. In Parkinson disease and dystonia, a
proposed mechanism of DBS is the suppression of
abnormal neuronal oscillations binding the basal ganglia–
thalamocortical network into a pathological functional
state.
41
Whether similar dysfunctional activity arises from
the perilesional area after stroke, as a result of maladap-
tive compensatory changes within the cortical motor net-
work or due to cortical deafferentation of the tonic
inhibitory basal ganglia input to the MLR, remains to be
elucidated in future studies.
Another aspect requiring additional research is the
optimal stimulation site within the MLR. The MLR is
primarily a functionally defined region at the
mesopontine junction; its anatomical substrate is not
fully characterized and still remains a matter of debate.
42
The MLR has been suggested to comprise noncholinergic
(ie, glutamatergic) cells that have been identified within
the lateral pontine tegmentum, confined medially by the
ventrolateral periaqueductal gray matter and laterally by
the PTg.
15
Electrical stimulation at various sites within
this region has elicited different forms of locomotor
behavior in various species, depending on the stimulation
amplitude. Because electrical pulses preferentially activate
myelinated fibers ortho- and antidromically, before small-
diameter fibers and cell bodies, it is difficult to discern
the anatomical substrate of a neurostimulation effect that
is often not local, but remote through modulation of
pathways rather than nuclei, even if the precise anatomi-
cal location of the electrode tip and the electrical field
distribution were known.
33
Moreover, a recent study of
pedunculopontine neurostimulation in a rat model of
Parkinson disease has cast doubt on a prominent locomo-
tor function of the PTg.
43,44
The overall effects of
pedunculopontine stimulation in patients have been dis-
appointing, apart from a group in which—by error—a
more lateral target site in the mesencephalon was cho-
sen,
45,46
possibly corresponding to the cuneiform nucleus
in man.
We are aware of several limitations in our study.
First, the number of animals (n 59 per group) is rela-
tively small and histological evaluation revealed a varia-
tion of placement of electrodes, which might explain the
variable stimulation outcome in this study. Second, the
precise location and extent of cortical motor representa-
tions for the left fore- and hindpaw have not been evalu-
ated electrophysiologically, which might contribute to a
relatively high variability of some postlesional gait param-
eters. However, such electrophysiological studies require a
fenestration of the skull, which—on the other hand—
results in a higher burden on animals and thus might
influence the locomotor behavior in the acute phase after
stroke. Third, the chosen coordinates for the photo-
thrombotic infarction also encompass somatosensory cor-
tical areas. Very small lesions result in gait impairments
that are too mild to be measured even with kinematic
analysis (CatWalk). Fourth, all MLR-HFS experiments
were conducted using wired neurostimulation by con-
necting the implanted electrode to an external stimulator.
However, because all experiments (especially the CatWalk
analysis) were performed by 2 investigators, the wire did
not substantially impact the MLR-HFS experiment; ani-
mals subjected to photothrombosis only (ie, not tethered
to a neurostimulator) showed no significant differences
in locomotor parameters when compared to rats connected
to the stimulator via electrical wire (sham-stimulation
Fluri et al: DBS for Gait Recovery
Month 2017 11

experiment, data not shown). It is of note that the resistance
of cerebral tissue varies widely in the phase shortly after
electrode implantation and induction of photothrombosis.
Thus, a stimulator with a large voltage compliance range is
needed to keep a constant current intensity to compare
experiments performed at different time points. The stimu-
lator used in this study is one of the few commercially avail-
able stimulator systems with this property and—to our
knowledge—there is no portable microstimulator for rats
having this feature.
In summary, and despite these uncertainties, this
study provides a novel framework for understanding
stroke-related gait disorder as a network dysfunction
amendable by neuromodulation, and thereby addresses
an as yet unmet clinical need in chronic stroke survivors.
Acknowledgment
This study was supported by the Deutsche Forschungsge-
meinschaft (Sonderforschungsbereich SFB 688) and by
the Interdisziplin€ares Zentrum f€ur Klinische Forschung,
University Hospital W€urzburg, W€urzburg, Germany.
We thank L. Frieß and A. Sauer for excellent technical
assistance; and D. Nock for critically proofreading the
manuscript.
Author Contributions
Conception and design of the experiments: F.F., J.V. Per-
formed the experiments: F.F. Analysis of data: all authors.
Processing of MRI data: G.A.H. Wrote the paper: F.F.,
J.V. Critical revision of the manuscript for important
intellectual content: all authors.
Potential Conflicts of Interest
Nothing to report.
References
1. van de Port IGL, Kwakkel G, Schepers VPM, Lindeman E. Predict-
ing mobility outcome one year after stroke: a prospective cohort
study. J Rehabil Med 2006;38:218–223.
2. Titianova EB, Pitk€
anen K, P€
a€
akk€
onen A, et al. Gait characteristics
and functional ambulation profile in patients with chronic unilateral
stroke. Am J Phys Med Rehabil 2003;82:778–786.
3. Partridge C, Edwards S. The bases of practice—neurological phys-
iotherapy. Physiother Res Int 1996;1:205–208.
4. Bobath B. Adult hemiplegia: evaluation and treatment. 3rd ed.
Oxford, UK: Oxford Butterworth-Heinemann, 1990.
5. Langhorne P, Coupar F, Pollock A. Motor recovery after stroke: a
systematic review. Lancet Neurol 2009;8:741–754.
6. Duysens J, Van de Crommert HWAA. Neural control of locomo-
tion; the central pattern generator from cats to humans. Gait Pos-
ture 1998;7:131–141.
7. Garcia-Rill E, Hyde J, Kezunovic N, et al. The physiology of the
pedunculopontine nucleus: implications for deep brain stimula-
tion. J Neural Transm 2014;122:225–235.
8. Shik ML, Severin FV, Orlovski K% GN. Control of walking and run-
ning by means of electric stimulation of the midbrain. Biofizika
1966;11:659–666.
9. Wenger N, Moraud EM, Gandar J, et al. Spatiotemporal neuro-
modulation therapies engaging muscle synergies improve motor
control after spinal cord injury. Nat Med 2016;22:138–145.
10. Guzzetta A, Pizzardi A, Belmonti V, et al. Hand movements at 3
months predict later hemiplegia in term infants with neonatal
cerebral infarction. Dev Med Child Neurol 2010;52:767–772.
11. Volkmann J, Allert N, Voges J, et al. Safety and efficacy of pallidal
or subthalamic nucleus stimulation in advanced PD. Neurology
2001;56:548–551.
12. Volkmann J, Wolters A, Kupsch A, et al. Pallidal deep brain stimulation
in patients with primary generalised or segmental dystonia: 5-year fol-
low-up of a randomisedtrial. Lancet Neurol 2012;11:1029–1038.
13. Fluri F, Schuhmann MK, Kleinschnitz C. Animal models of ischemic
stroke and their application in clinical research. Drug Des Devel
Ther 2015;9:3445–3454.
14. Jahn K, Deutschl€
ander A, Stephan T, et al. Supraspinal locomotor con-
trol in quadrupeds and humans. Prog Brain Res 2008;171:353–362.
15. Sherman D, Fuller PM, Marcus J, et al. Anatomical location of the
mesencephalic locomotor region and its possible role in locomotion,
posture, cataplexy,and parkinsonism. Front Neurol 2015;6:140.
16. Skinner RD, Garcia-Rill E. The mesencephalic locomotor region
(MLR) in the rat. Brain Res 1984;323:385–389.
17. Cabelguen J-M, Bourcier-Lucas C, Dubuc R. Bimodal locomotion
elicited by electrical stimulation of the midbrain in the salamander
Notophthalmus viridescens. J Neurosci 2003;23:2434–2439.
18. Eidelberg E, Walden JG, Nguyen LH. Locomotor control in
macaque monkeys. Brain 1981;104:647–663.
19. Golestanirad L, Elahi B, Graham SJ, et al. Efficacy and safety of
pedunculopontine nuclei (PPN) deep brain stimulation in the treat-
ment of gait disorders: a meta-analysis of clinical studies. Can J
Neurol Sci 2016;43:120–126.
20. Nosko D, Ferraye MU, Fraix V, et al. Low-frequency versus high-
frequency stimulation of the pedunculopontine nucleus area in
Parkinson’s disease: a randomised controlled trial. J Neurol Neu-
rosurg Psychiatry 2015;86:674–679.
21. Welter M-L, Demain A, Ewenczyk C, et al. PPNa-DBS for gait and
balance disorders in Parkinson’s disease: a double-blind, rando-
mised study. J Neurol 2015;262:1515–1525.
22. Lindau NT, B€
anninger BJ, Gullo M, et al. Rewiring of the cortico-
spinal tract in the adult rat after unilateral stroke and anti-Nogo-A
therapy. Brain 2014;137(pt 3):739–756.
23. Paxinos G, Watson C. The rat brain in stereotaxic coordinates. 6th
ed. Amsterdam, the Netherlands: Academic Press Elsevier, 2007.
24. Fluri F, Bieber M, Volkmann J, Kleinschnitz C. Microelectrode
guided implantation of electrodes into the subthalamic nucleus of
rats for long-term deep brain stimulation. J Vis Exp 2015;(104).
25. Bachmann LC, Matis A, Lindau NT, et al. Deep brain stimulation of
the midbrain locomotor region improves paretic hindlimb function
after spinal cord injury in rats. Sci Transl Med 2013;5:208ra146.
26. Matsumura M, Nambu A, Yamaji Y, et al. Organization of somatic
motor inputs from the frontal lobe to the pedunculopontine teg-
mental nucleus in the macaque monkey. Neuroscience 2000;98:
97–110.
27. Hamers FPT, Lankhorst AJ, van Laar TJ, et al. Automated quanti-
tative gait analysis during overground locomotion in the rat: its
application to spinal cord contusion and transection injuries.
J Neurotrauma 2001;18:187–201.
28. Encarnacion A, Horie N, Keren-Gill H, et al. Long-term behavioral
assessment of function in an experimental model for ischemic
stroke. J Neurosci Methods 2011;196:247–257.
ANNALS of Neurology
12 Volume 00, No. 0

29. Parkkinen S, Ortega FJ, Kuptsova K, et al. Gait impairment in a
rat model of focal cerebral ischemia. Stroke Res Treat 2013;2013:
410972.
30. Feeney DM, Gonzalez A, Law WA. Amphetamine, haloperidol,
and experience interact to affect rate of recovery after motor cor-
tex injury. Science 1982;217:855–857.
31. Neckel ND, Dai H, Bregman BS. Quantifying changes following
spinal cord injury with velocity dependent locomotor measures.
J Neurosci Methods 2013;214:27–36.
32. Koopmans GC, Brans M, G
omez-Pinilla F, et al. Circulating
insulin-like growth factor I and functional recovery from spinal
cord injury under enriched housing conditions. Eur J Neurosci
2006;23:1035–1046.
33. Ranck JB Jr. Which elements are excited in electrical stimulation
of mammalian central nervous system: a review. Brain Res 1975;
98:417–440.
34. la Fouge` re C, Zwergal A, Rominger A, et al. Real versus imagined
locomotion: a [18F]-FDG PET-fMRI comparison. Neuroimage
2010;50:1589–1598.
35. Esposito MS, Capelli P, Arber S. Brainstem nucleus MdV mediates
skilled forelimb motor tasks. Nature 2014;508:351–356.
36. Martinez-Gonzalez C, Bolam JP, Mena-Segovia J. Topographical
organization of the pedunculopontine nucleus. Front Neuroanat
2011;5:22.
37. Perry J, Garrett M, Gronley JK, Mulroy SJ. Classification of
walking handicap in the stroke population. Stroke 1995;26:982–
989.
38. Michael KM, Allen JK, Macko RF. Reduced ambulatory activity
after stroke: the role of balance, gait, and cardiovascular fitness.
Arch Phys Med Rehabil 2005;86:1552–1556.
39. Roseberry TK, Lee AM, Lalive AL, et al. Cell-type-specific control
of brainstem locomotor circuits by basal ganglia. Cell 2016;164:
526–537.
40. Takakusaki K, Habaguchi T, Ohtinata-Sugimoto J, et al. Basal gan-
glia efferents to the brainstem centers controlling postural muscle
tone and locomotion: a new concept for understanding motor dis-
orders in basal ganglia dysfunction. Neuroscience 2003;119:293–
308.
41. Wichmann T, DeLong MR. Deep brain stimulation for movement
disorders of basal ganglia origin: restoring function or functional-
ity? Neurotherapeutics 2016;13:264–283.
42. Ryczko D, Dubuc R. The Multifunctional Mesencephalic Locomotor
Region. Curr Pharm Des 2013;19:4448–4470.
43. Gut NK, Winn P. The pedunculopontine tegmental nucleus—a
functional hypothesis from the comparative literature. Mov Disord
2016;31:615–624.
44. Gut NK, Winn P. Deep brain stimulation of different pedunculo-
pontine targets in a novel rodent model of parkinsonism.
J Neurosci 2015;35:4792–4803.
45. Stefani A, Lozano AM, Peppe A, et al. Bilateral deep brain stimu-
lation of the pedunculopontine and subthalamic nuclei in severe
Parkinson’s disease. Brain 2007;130:1596–1607.
46. Zrinzo L, Zrinzo LV, Hariz M. The pedunculopontine and peripe-
duncular nuclei: a tale of two structures. Brain 2007;130:e73.
Fluri et al: DBS for Gait Recovery
Month 2017 13