Gait & Posture 39 (2014) 563569

Contents lists available at ScienceDirect

Gait & Posture

journal homepage: www.elsevier.com/locate/gaitpost

Functional recovery and rehabilitation of postural impairment and gait

ataxia in patients with acute cerebellar stroke
Uta Bultmann a,1, Daniela Pierscianek b,1, Elke R. Gizewski c, Beate Schoch d,
Nicole Fritsche a, Dagmar Timmann a, Matthias Maschke e, Markus Frings a,*
a
Department of Neurology, University of Duisburg-Essen, Hufelandstr. 55, 45122 Essen, Germany
b
Department of Neurosurgery, University of Duisburg-Essen, Hufelandstr. 55, 45122 Essen, Germany
c
Clinic for Neuroradiology, Medical University Innsbruck, Anichstr. 35, Innsbruck, Austria
d
Department of Neurosurgery, Stiftungsklinikum Mittelrhein Koblenz, Johannes-Muller-Str. 7, 56068 Koblenz, Germany
e
Department of Neurology, Bruderkrankenhaus, Nordallee 1, 54292 Trier, Germany

A R T I C L E I N F O A B S T R A C T

Article history: Studies about recovery from cerebellar stroke are rare. The present study assessed motor decits in the
Received 15 April 2013 acute phase after isolated cerebellar stroke focusing on postural impairment and gait ataxia and outlines
Received in revised form 12 August 2013 the role of lesion site on motor outcome, the course of recovery and the effect of treadmill training. 23
Accepted 13 September 2013
patients with acute and isolated cerebellar infarction participated. Decits were quantied by ataxia
scores and dynamic posturography in the acute phase and in a follow up after 2 weeks and 3 months. MRI
Keywords: data were obtained to correlate lesion site with motor performance. Half of the patients that gave
Cerebellar infarction
informed consent and walked independently underwent a 2-week treadmill training with increasing
Ataxia
Brain imaging
velocity. In the acute phase patients showed a mild to severe ataxia with a worse performance in patients
Treadmill training with infarction of the superior in comparison to the posterior inferior cerebellar artery. However, after 3
Voxel based lesion symptom mapping months differences between vascular territories were no longer signicant. MRI data showed that
patients with larger infarct volumes had a signicantly more severe ataxia. In patients with ataxia of
stance, gait and lower limbs lesions were more common in cerebellar lobules IV to VI. After 3 months a
mild ataxia in lower limbs and gait, especially in gait speed persisted. Because postural impairment had
fully recovered, remaining gait ataxia was likely related to incoordination of lower limbs. Treadmill
training did not show signicant effects. Future studies are needed to investigate whether intensive
coordinative training is of benet in patients with cerebellar stroke.
2013 Published by Elsevier B.V.

1. Introduction cerebellar artery (SCA) supplies the superior cerebellum from

Only 23% of ischaemic strokes are isolated infarctions of the
cerebellum [1,2]. In contrast to brainstem or neocortical infarction,
symptoms of cerebellar infarction are often non-specic and
moderate, like nausea, dizziness or headache. As a challenge in
diagnosis the characteristic limb ataxia only occurs in 40% [1,3,4].
Previous studies described a functional compartmentalization of
the human cerebellum, indicating that various areas within the
cerebellum have different functions corresponding to their afferent
and efferent connections [1,3,5]. Therefore, infarctions of different
areas in the cerebellum lead to distinct symptoms. The superior
* Corresponding author at: Department of Neurology, University of Duisburg-
Essen, Hufelandstrasse 55, D-45122 Essen, Germany. Tel.: +49 201 723 2461;
fax: +49 201 723 5901.
E-mail addresses: markus.frings@uni-duisburg-essen.de,
markus.frings@uni-due.de (M. Frings).
1
These authors contributed equally to this work.
hemispheral and vermal lobule I to parts of lobule VII and most of
cerebellar nuclei. Infarction of the SCA causes pronounced limb
ataxia combined with gait ataxia when a cerebellar hemisphere is
affected and dysarthria when paravermal regions are affected. The
posterior inferior cerebellar artery (PICA) supplies the more
inferior parts of the cerebellum and parts of the dentate nucleus.
Infarction of the PICA leads to gait and postural instability,
nystagmus and vertigo [1,5]. In general, patients with infarction of
the SCA seem to have a poorer outcome in motor functions than
patients with infarction of the PICA [2,6,7].
The present study focused on functional recovery of postural
impairment, gait and limb ataxia after acute cerebellar stroke.
Recent lesion studies using voxel based lesion symptom mapping
showed a signicant correlation of lower limb ataxia with lesions
of vermal and hemispheral lobules III to IV and of ataxia of posture
and gait with lesions of vermal lobules II to IV [8]. The territory of
the SCA includes all these regions. Therefore, it is one hypothesis of
the present study, that the impairment of posture and gait may be

0966-6362/$ see front matter 2013 Published by Elsevier B.V.

http://dx.doi.org/10.1016/j.gaitpost.2013.09.011
564 U. Bultmann et al. / Gait & Posture 39 (2014) 563569

pronounced in patients with SCA infarction. According to a Table 1

Patient characteristics, type, and side of stroke and clinical ataxia score.
previous study in patients with cerebellar tumours, a poorer
outcome in patients with an involvement of cerebellar nuclei was Patient Age (yr) Gender Stroke Side International Cooperative
expected, too [9]. A further expectation was a good recovery from Ataxia Rating Scale, total score
symptoms according to a prior study focusing on ataxia of the Session 1 Session 2 Session 3
upper limbs in the same patient group as in the present study [10]. 01 77 M SCA Right 18 2 2
Standardized ataxia scores and dynamic posturography were used 02 59 M SCA Left 8
to quantify the impairments. In order to correlate the motor 03 81 M SCA Left 37
decits with the lesion site voxel based lesion symptom mapping 04 71 M SCA Right 22 11 3
05 55 M SCA Left 6 2 1
was applied. There are several studies which investigate the
06 43 M PICA Right 6 4 4
rehabilitation of patients suffering from neocortical infarction [11 07 44 M PICA Left 2 0 0
13]. In contrast, rehabilitation programmes for patients after 08 49 M PICA Right 3 0 0
cerebellar stroke are rare. So far the effects of special rehabilitation 09 55 M SCA Right 42 23 16
programmes like balance training, coordination training or 10 48 F SCA Right 5 2 0
11 46 F PICA Left 16 7 2
treadmill training are only examined in patients with chronic
12 83 F SCA Left 46
cerebellar degeneration [1417]. Therefore, in another part of the 13 60 M PICA Left 14 9 8
present study, according to a well-established approach in patients 14 61 M SCA Right 62 43 25
with hemiparetic stroke [12], an effect of treadmill training with 15 75 F PICA Right 11 6 0
16 59 M PICA Right 9 9 5
increasing velocity on ataxia of gait and posture should be
17 82 F SCA Left 20 20 18
examined. 18 47 M PICA Left 5 4 0
19 61 M PICA Right 13 12 4
2. Subjects and methods 20 48 F PICA Right 6
21 42 F PICA Right 9
22 58 F PICA Left 6
2.1. Subjects 23 44 F PICA Left 6

A total of 23 patients with acute, isolated cerebellar stroke (6
females and 17 males, mean age 58.6  13.5 yrs) and 13 healthy
age-matched controls (4 females and 9 males, mean age
59.3  10.5 yrs) were enrolled in this study. In detail, ten patients
presented with a stroke in the territory of the superior cerebellar
artery (SCA) and 13 patients had a stroke of the posterior inferior
cerebellar artery (PICA). On initial presentation, eleven of 23 patients
showed cerebellar lesions that included the cerebellar nuclei. Only
patients with an isolated cerebellar stroke were included in the study,
veried by an initial MRI scan. Mean duration from diagnosis of stroke
until enrolment in the study was 12.4  11.7 (range 141) days. Most
of the patients and control subjects of the present study have already
been examined in a recent study of recovery of upper limb ataxia [10].
In a multicentre-study patients were recruited from the university
hospital in Essen and ve other neurological hospitals in the
surroundings. Fig. 1 gives a detailed explanation of the relationship
of patients of the acute stage, of the follow-up investigation and
treadmill training. A detailed description of the patients character-
istics is given in Table 1. All subjects gave their written informed
consent prior to their inclusion in the study. The study was approved
by the local ethics committee and has therefore been performed in
accordance with the ethical standards laid down in the 1964
Declaration of Helsinki and its later amendments.
F indicates female; M, male; PICA, posterior inferior cerebellar artery; SCA, superior
cerebellar artery.
2.2. International Cooperative Ataxia Rating Scale (ICARS)
Cerebellar symptoms were quantied using the International
Cooperative Ataxia Rating Scale (ICARS) of the World Federation of
Neurology [18]. ICARS ranges from 0 (no ataxia) to 100 (strongest
ataxia) and comprises four main clinical core symptoms: posture
and gait disturbances, kinetic functions, speech disorders and
oculomotor disorders. As it has been reported previously the
scores were subdivided into gait (max. score 12), posture (max.
score 18) and lower limb (max. score 16) to describe more in
detail the relevant ataxia for the applied methods [9]. Additionally,
the gait score was subdivided into the parameters walking
capacities and gait speed.
2.3. Dynamic posturography
Postural stability was examined by dynamic posturography
using the Sensory Organization Test (SOT) of the EquiTest1 system
(NeuroCom Inc., Portland, OR) and might be regarded as an
advanced version of the Romberg Test. The application of SOT for

Fig. 1. Flow chart with an explanation of the relationship of patients of the acute stage, of the follow-up investigations and treadmill training and reasons for the loss of
patients.
 U. Bultmann et al. / Gait & Posture 39 (2014) 563569 565

the assessment of postural stability in cerebellar diseases has been
described in detail previously [9,19]. The SOT measures the body
sway and comprises six different conditions with three repetitions
for each condition. For analysis, the total sway area of the centre of
gravity was calculated based on changes in ground reaction force,
using custom-made algorithms.
2.4. MR imaging
All patients received a cranial MRI using a Siemens Sonata 1.5 T
to assess the location and size of the stroke and to exclude
extracerebellar lesions. Methods of acquisition and processing of
MRI have been described in detail previously [10].
Additionally, voxel-based lesion symptom mapping (VLSM)
of the cerebellum was performed [20]. This analysis calculates
for each voxel the statistical probability that a lesion in this
voxel is accompanied by an abnormal performance using the
permutated BrunnerMunzel rank order statistic. The statistical
maps were superimposed on MRI images of a healthy brain
(Fig. 2).
2.5. Rehabilitation treatment
In order to test the usefulness and efciency of a rehabilitation
treatment during the acute phase after cerebellar stroke, patients
were randomized and half of the patients underwent treadmill
training during the rst 2 weeks after enrollment in the study.
Patients that were not able to walk alone were excluded from
treadmill training to prevent falls. This training took place
additionally to a classical physiotherapy after stroke according
to the Bobath concept. The treadmill training was carried out by an
experienced therapist (N.F.) and lasted 30 min per day. Velocity
during the training was varied and progressively increased
according to the patients ability.
2.6. Statistical analysis
Statistical analysis of the data was performed using SPSS 20.0.0
(SPSS Inc., Chicago, IL, USA). Univariate ANOVA were calculated to
assess differences between patients and controls and in post hoc
comparisons between the different patients groups (SCA vs. PICA
infarction, affected vs. unaffected cerebellar nuclei, left vs. right
side) at the acute stage. Additional repeated-measures ANOVA
were calculated over three points of time, i.e. the acute stage, after
2 weeks and 3 months, to assess differences of recovery between
groups (block by group interaction effect). Because only 16 of
initially 23 patients were available for follow-up, a univariate
ANOVA for this group of patients was repeated for the acute stage
and calculated for the last examination after 3 months. Level of
signicance was set to P < 0.05.

Fig. 2. Superimposed acute lesions (a) of all 23 patients with cerebellar infarctions. Patients lesions are superimposed on a MRI of a healthy control subject and right-sided
lesions were ipped to the left. The number of overlapping lesions is indicated by the colour scale. There are two centres of lesion overlap, in the upper cerebellum,
representing the SCA territory, and in the lower cerebellum according to PICA territory. Voxel-based lesion-symptom mapping (b and c) of all 23 patients. Z-values are
indicated by the colour scale. Z-values between 2.5 and 3.2 (corresponding to P-values between 0.05 and 0.01) are shown. Analysis revealed that an area spanning vermal
lobules IV to the hemispheric lobule VI including the dentate nuclei was signicantly associated with a higher total ICARS score (b). For the ICARS subscores gait (c),
posture and lower limb, analysis showed the highest t-values in the same area. (For interpretation of the references to colour in this gure legend, the reader is referred to
the web version of the article.)
566 U. Bultmann et al. / Gait & Posture 39 (2014) 563569

3. Results 3.1.2. ICARS across the 3 months

3.1. Clinical scores
3.1.1. ICARS in the acute stage
For the initial quantication of cerebellar symptoms using the
ICARS score, 23 patients and 13 healthy controls were investigated.
Univariate analysis of variance (ANOVA) for the initial investiga-
tion revealed for the total ICARS score, as also for the subscore
gait, posture and lower limb a signicant difference between
patients and controls (P-values <0.01), as shown in Fig. 3. Of the 23
patients, ten patients had a stroke in the territory of the SCA and 13
patients in the territory of the PICA. Patients with a SCA infarction
had signicantly higher total ICARS score (P = 0.003) as also ICARS
subscores (P-values <0.03) compared to patients with a PICA
infarction (Fig. 3). Twelve patients had infarctions on the right side
of the cerebellum and eleven patients had left-sided strokes. No
signicant difference regarding the infarctions side could be
detected for ICARS total score and subscores (P > 0.5). Eleven
patients showed an affection of cerebellar nuclei on initial MRI
scan. The comparison of ICARS total score and subscores (P-values
>0.09) between patients with and without affection of cerebellar
nuclei did not reveal a signicant difference. Patients with larger
infarct volumes had a signicantly higher total ICARS score
(P = 0.03, bivariate correlation analysis), whereas for the ICARS
subscores there was no signicant correlation present (P-values
>0.09).
Sixteen of 23 patients and 13 healthy controls were investigat-
ed additionally after 2 weeks and after 3 months (Fig. 1). Among
those 16 patients, seven patients had a SCA infarction and nine
patients a PICA infarction. To make sure that differences at the
beginning were signicant in this smaller follow-up group, too,
ANOVA of the acute stage was repeated. For the initial point of
time, the difference for the total ICARS score between patients with
SCA infarction and PICA infarctions was signicant for this group of
16 patients and 13 controls, too (P = 0.04), but not for ICARS
subscores (P-values >0.06).
A repeated analysis of variance over 3 months revealed only a
tendency to higher total ICARS scores in patients with SCA
infarction (P = 0.06). However, there was a signicant block-effect
(P < 0.001) and over time, patients with SCA infarction improved
signicantly more than patients with PICA infarction (block by
group-effect; P = 0.02). For the ICARS subscores, no group or block
by group interaction effect could be shown (P-values >0.08).
In another repeated analysis of variance over 3 months no
signicant differences were found for the total ICARS score and
subscores between patients with right- and left-sided infarctions
(P-values >0.2) and between patients with and without involve-
ment of cerebellar nuclei (P-values >0.8).
3.1.3. ICARS after 3 months
After 3 months, there were signicant differences between
patients (N = 16) and healthy controls (N = 13) only for the total

Fig. 3. Comparison of the total ICARS score and its subscores for gait, posture and lower limb between patients and controls, as also between patient subgroups
regarding the stroke territory (SCA, PICA), side of infarction and involvement of cerebellar nuclei. Mean score in percent with standard error of the mean is shown. ICARS total
score and all subscores were higher in patients with SCA than with PICA infarction.
 U. Bultmann et al. / Gait & Posture 39 (2014) 563569
ICARS score, ICARS gait and ICARS lower limb (P = 0.04,
P = 0.02, P = 0.04, respectively), but there was no difference for
ICARS posture (P = 0.12). Depending on the vascular territory, the
side of infarction or involvement of cerebellar nuclei there was no
signicant difference after 3 months (P-values >0.08).
3.1.4. ICARS gait analysis
We performed an additional statistical analysis between
patients and controls comparing walking capacities and gait speed
as part of the gait score of ICARS. In a univariate ANOVA at the
acute stage the group of 23 patients showed worse walking
capacities than the group of control subjects (group effect,
P = 0.006). Comparison of the follow-up group of 16 patients with
controls showed a signicant group effect at the acute stage, too
(P = 0.01). After 3 months differences were no longer signicant
(P = 0.2). In an ANOVA with repeated measures group effect was
P = 0.02, block effect P = 0.007 and block by group interaction effect
(P = 0.007). Analysis of gait speed at the acute stage showed a
signicant group effect for all 23 patients (P < 0.001) and for the 16
patients of the follow-up group (P = 0.01) compared to controls.
After 3 months the difference between patients and controls was
still signicant (P = 0.002). An ANOVA with repeated measures
over the 3 months showed a signicant group effect (P = 0.001),
block effect (P = 0.02) and block by group effect (P = 0.02).
3.2. Dynamic posturography
The ten patients that took part in the rehabilitation training and
twelve controls were included in the analysis of dynamic
posturography (Fig. 4). In the repeated-measures ANOVA across
the three points in time, there was a signicant greater sway area
for patients than for controls for all six conditions together
(P = 0.04). There was a signicant block-effect with a general
decrease of the sway area during the 3 months (P < 0.001). The
decrease in sway area was greater for patients than for controls
over the 3 months (P = 0.04, block by group-effect). A post hoc
analysis revealed a signicant difference between patients and
controls for condition 6 (P = 0.04), where the subjects performance
depends on the vestibular system. A repeated-measures ANOVA
for all six conditions as also for each condition separately did not
show signicant differences between patients with SCA infarction
compared to patients with PICA infarction, between patients with
left- and right-sided infarctions or between patients with and
without involvement of cerebellar nuclei (P-values >0.07).
3.3. MR imaging
The mean infarction volume on the MRI for all 23 patients was
7.01  7.0 cm3 (range 0.333.7 cm3). Further analysis revealed two
centres of lesion overlap: In the upper cerebellum, representing the
SCA territory, and in the lower cerebellum according to PICA territory
(Fig. 2). The lesion in the upper cerebellar part was assigned to lobules
IV/Crus I, whereas the lower cerebellar lesion site was identied as
lobules VII and VIII. Voxel-based lesion-symptom mapping revealed
that an area spanning vermal lobules IV to the hemispheric lobule VI
including the dentate nuclei was signicantly associated with a
higher total ICARS score and subscores.
3.4. Rehabilitation training
When comparing the two groups regarding the total ICARS
score and its subscores (treadmill-trained group N = 5 vs. no
treadmill training N = 5), there were no signicant differences (P-
values >0.85) on the initial exam. Over the 3 months, there was a
signicant block-effect (P < 0.03), but no block by group interac-
tion effect for the total ICARS scores and its subscores (P-values
567
>0.99). After 3 months, no signicant difference was detected
between patients who received treadmill training and those
without treadmill training for total ICARS score (P = 0.59) and for
the ICARS subscores (P-values >0.35).
4. Discussion
In the acute phase after an isolated cerebellar stroke patients
had a mild to severe postural impairment and ataxia of gait and
lower limbs, which was shown in the ataxia scores and in the
dynamic posturography. Patients with an infarction of the SCA
were more impaired than patients with an infarction of the PICA. A
comparison of ataxia in patients with right- and left-sided
infarctions and with and without an involvement of the cerebellar
nuclei did not show signicant differences. Secondly, the recovery
of these motor decits within the rst 3 months after stroke was
high. No balance decits could be detected after 3 months, while
mild decits in lower limbs and gait, especially in gait speed
persisted. Thirdly, by using voxel based lesion symptom mapping a
correlation between the ataxia scores and a lesion in lobules IV to
VI in the upper cerebellum was shown. Furthermore, patients with
larger infarct volumes had a signicantly higher total ICARS score,
i.e. a more severe ataxia. Finally, additional treadmill training did
not lead to a further improvement of the motor decits.
Because of the missing decit in the ICARS subscore posture
and in posturography in the present study after 3 months, and
persistent decits in the ICARS subscores gait and lower limb,
we suppose a relationship between the ataxia of lower limb and
gait disturbances. These results are in line with previous ndings
that disordered intra limb coordination with increased timing
variability seemed to contribute to walking decits in patients
with isolated cerebellar stroke [21]. Of note, this does not exclude
the contribution of balance decits [22,23]. Comparing patients
with an infarction of the SCA or the PICA in the initial phase after
stroke, patients with infarction of the SCA performed worse in the
total ICARS score as well as in all subscores. However, in the course
of the study, no differences were detected anymore. Accordingly,
within the 3 months, patients with an infarction of the SCA
improved signicantly better than patients with an infarction of
the PICA. Most relevant structures for limb movements and gait are
located in the superior part of the cerebellum which is supplied by
the SCA [8]. Therefore, patients with SCA infarction show more
motor impairment than patients with PICA infarction at the acute
stage. The main reason for the lack of a difference between vascular
territories after 3 months may be an adjustment of both groups by
the good spontaneous recovery of patients with cerebellar stroke
in general. As a reason for this good recovery an activation of
contralateral cerebellar and neocortical areas is discussed [7].
In the present study, no signicant difference in the functional
outcome between patients with affection and without an affection
of the cerebellar nuclei was shown in contrast to previous studies
[8,9,19]. One reason for the missing expected association may be
the small sample size. Furthermore, oedema in the acute phase
after the stroke has to be differentiated from permanent lesions
and oedema may have pretended an involvement of cerebellar
nuclei in some patients.
In hemiparetic patients, treadmill training was performed with
increasing speed during training adapting to the personal
capability of the patients as a kind of forced induced strategy
[12]. In the present study, treadmill velocity was increased during
training, too. The nding of a persistent reduced gait speed in
patients after cerebellar stroke justies the attempt of a training
with increasing speed. However, no effect of treadmill training
could be shown in the present study. A possible reason for the lack
of benets from the treadmill training could be, that decreased
velocity is not a cerebellar induced abnormality but a safety or
568 U. Bultmann et al. / Gait & Posture 39 (2014) 563569

Fig. 4. Comparison of mean sway area in cm2 between patients and controls for the six conditions of the Sensory Organization Test (SOT). In condition 1 balance function is
measured under normal sensory conditions and in condition 2 with eyes closed. During the other conditions the sensory information is varied by rotating the visual surround
(condition 3) and/or the platform on which the subject is standing (condition 4) in a sway-referenced matter. Sway-referenced corresponds to anteriorposterior motion of
the visual surround or the platform that is coupled to the subjects own sway. In condition 5 the subjects eyes are closed and the platform is sway-referenced whereas in
condition 6 both parameters are sway-referenced. Mean score of three repetitions in percent with standard error of the mean and the absolute number of falls is shown for the
acute phase (1), after 2 weeks (2) and 3 months (3). The decrease in sway area was greater for patients than for controls over the 3 months.

compensation strategy ot the patients. Another reason may be, that
the spontaneous recovery after acute cerebellar stroke was high. As
a further explanation, the treadmill training of the present study
might have been not challenging enough for patients with only
mild ataxia. Because of the persisting limb ataxia in the chronic
state the improvement of the intra limb coordination seems to be
important. Thereby, the efciency of a treadmill training may be
enhanced by increasing the degree of difculty of the training,
possibly, by increasing the declination of the treadmill or by
introducing obstacles. In the future, different therapies need to be
compared in randomized controlled trials.
Conicts of Interest
The authors declare that they have no conict of interest in this
study.
 U. Bultmann et al. / Gait & Posture 39 (2014) 563569
Acknowledgement
The study was supported by a grant from the German Federal
Ministry for Science and Research (BMBF 01 GA 0508).
References
[1] Amarenco P. The spectrum of cerebellar infarction. Neurology 1991;41:9739.
[2] Tohgi H, Takahashi S, Chiba K, Hirata Y. Cerebellar infarction. Clinical and
neuroimaging analysis in 293 patients. The Tohoku Cerebellar Infarction Study
Group. Stroke 1993;24:1697701.
[3] Amareno P. Les infarctus du cervelet et leurs mecanismes. Rev Neurol
1993;149:72748.
[4] Edlow JA, Newman-Toker DE, Savitz SI. Diagnosis and initial management of
cerebellar infarction. Lancet Neurol 2008;7:95164.
[5] Dichgans J. Clinical symptoms of cerebellar dysfunction and their topodiag-
nostical signicance. Hum Neurobiol 1984;2:26979.
[6] Kelly PJ, Stein J, Shafqat S, Eskey C, Doherty D, Chang Y, et al. Functional
recovery after rehabilitation for cerebellar stroke. Stroke 2001;32:5304.
[7] Kinomoto K, Takayama Y, Watanabe T, Kawasaki T, Onishi K, Yagi H, et al. The
mechanisms of recovery from cerebellar infarction: an fMRI study. Neurore-
port 2003;14:16715.
[8] Schoch B, Dimitrova A, Gizewski ER, Timmann D. Functional localisation in the
human cerebellum based on voxelwise statistical analysis: a study of 90
patients. Neuroimage 2006;30:3651.
[9] Schoch B, Konczak J, Dimitrova A, Gizewski ER, Wieland R, Timmann D. Impact
of surgery and adjuvant therapy on balance function in children and adoles-
cents with cerebellar tumors. Neuropediatrics 2006;37:3508.
[10] Konczak J, Pierscianek D, Hirsiger S, Bultmann U, Schoch B, Gizewski ER, et al.
Recovery of upper limb function after cerebellar stroke: lesion symptom
mapping and arm kinematics. Stroke 2010;41:2191200.
[11] Hesse S, Bertelt C, Jahnke MT, Schaffrin A, Baake P, Malezic M, et al. Treadmill
training with partial body weight support compared with physiotherapy in
nonambulatory hemiparetic patients. Stroke 1995;26:97681.
569
[12] Pohl M, Mehrholz J, Ritschel C, Ruckriem S. Speed-dependent treadmill
training in ambulatory hemiparetic stroke patients: a randomized controlled
trial. Stroke 2002;33:5538.
[13] Visintin M, Barbeau H, Korner-Bitensky N, Mayo NE. A new approach to retrain
gait in stroke patients through body weight support and treadmill stimulation.
Stroke 1998;29:11228.
[14] Balliet R, Harbst KB, Kim D, Stewart RV. Retraining of functional gait through
the reduction of upper extremity weight-bearing in chronic cerebellar ataxia.
Int Rehabil Med 1987;8:48153.
[15] Cernak K, Stevens V, Price R, Shumway-Cook A. Locomotor training
using body-weight support treadmill in conjunction with ongoing physical
therapy in a child with severe cerebellar ataxia. Phys Ther 2008;88:
8898.
[16] Gill-Body K, Pobat R, Parker SW, Krebs DE. Rehabilitation of balance in two
patients with cerebellar dysfunction. Phys Ther 1997;77:53452.
[17] Ilg W, Synofzik M, Brotz D, Burkard S, Giese MA, Schols L. Intensive coordi-
native training improves motor performance in degenerative cerebellar dis-
ease. Neurology 2009;73:182330.
[18] Trouillas P, Takayanagi T, Hallett M, Currier RD, Subramony SH, Wessel K,
et al. International Cooperative Ataxia Rating Scale for pharmacological
assessment of the cerebellar syndrome. The Ataxia Neuropharmacology
Committee of the World Federation of Neurology. J Neurol Sci 1997;
12:20511.
[19] Konczak J, Schoch B, Dimitrova A, Gizewski E, Timmann D. Functional recovery
of children and adolescents after cerebellar tumor resection. Brain 2005;
128:142841.
[20] Bates E, Wilson SM, Saygin AP, Dick F, Sereno MI, Knight RT, et al. Voxel-based
lesion symptom mapping. Nat Neurosci 2003;6:44850.
[21] Ilg W, Golla H, Thier P, Giese MA. Specic inuences of cerebellar dysfunction
on gait. Brain 2007;130:78698.
[22] Morton S, Bastian AJ. Relative contributions of balance and voluntary leg-
coordination decits to cerebellar gait ataxia. J Neurophysiol 2003;89:
184456.
[23] Stolze H, Klebe S, Petersen G, Raethjen J, Wenzelburger R, Witt K, et al. Typical
features of cerebellar ataxic gait. Neurol Neurosurg Psychiatry 2002;73:
3102.