Journal of the Neurological Sciences 377 (2017) 25–30

Contents lists available at ScienceDirect

Journal of the Neurological Sciences

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

Robotic gait training in multiple sclerosis rehabilitation: Can virtual

reality make the difference? Findings from a randomized controlled trial
Rocco Salvatore Calabrò ⁎,1, Margherita Russo 1, Antonino Naro, Rosaria De Luca, Antonino Leo,
Provvidenza Tomasello, Francesco Molonia, Vincenzo Dattola, Alessia Bramanti, Placido Bramanti
IRCCS Centro Neurolesi “Bonino-Pulejo”, C.da Casazza SS. 113, Messina, Italy

a r t i c l e i n f o a b s t r a c t

Article history: Gait, coordination, and balance may be severely compromised in patients with multiple sclerosis (MS), with con-
Received 13 January 2017 siderable consequences on the patient's daily living activities, psychological status and quality of life. For this rea-
Received in revised form 24 February 2017 son, MS patients may benefit from robotic-rehabilitation and virtual reality training sessions. Aim of the present
Accepted 28 March 2017
study was to assess the efficacy of robot-assisted gait training (RAGT) equipped with virtual reality (VR) system
Available online 29 March 2017
in MS patients with walking disabilities (EDSS 4.0 to 5.5) as compared to RAGT without VR. We enrolled 40 pa-
Keywords:
tients (randomized into two groups) undergoing forty RAGT ± VR sessions over eight weeks. All the patients
Multiple sclerosis were assessed at baseline and at the end of the treatment by using specific scales. Effect sizes were very small
Lokomat and non-significant between the groups for Berg Balance Scale (− 0.019, CI95% − 2.403 to 2.365) and TUG
Virtual reality (−0.064, 95%CI −0.408 to 0.536) favoring RAGT + VR. Effects were moderate-to-large and significant for posi-
Robotic rehabilitation tive attitude (−0.505, 95%CI −3.615 to 2.604) and problem-solving (−0.905, 95%CI −2.113 to 0.302) sub-items
of Coping Orientation to Problem Experienced, thus largely favoring RAGT + VR. Our findings show that RAGT
combined with VR is an effective therapeutic option in MS patients with walking disability as compared to
RAGT without VR. We may hypothesize that VR may strengthen RAGT thanks to the entrainment of different
brain areas involved in motor panning and learning.
© 2017 Elsevier B.V. All rights reserved.

1. Introduction
Multiple sclerosis (MS) is the second most common cause of neuro-
logical disability in adults aged 18–50 [1]. Ambulation may be compro-
mised early, even in people with mild disability, independently of the
relapsing or progressive onset [2]. Specific gait abnormalities in ambula-
tory MS patients include reduced velocity and stride length, increased
double-limb support time, and gait asymmetries.
Currently, few studies have shown some positive effects of human-
ized monoclonal antibody therapy and 4.aminopyridine in improving
walking speed in MS patients [3–4]. On the other hand, there is growing
evidence that subjects with progressive or relapsing-remitting MS [5–6]
may benefit from rehabilitation sessions.
In the last few years, novel locomotor devices combining body
weight-supported treadmill training (BWSTT) have been developed to
enhance gait recovery following central nervous system injury [7,8].
Further, robotic-assisted gait training (RAGT), including Lokomat, have
been developed to facilitate the delivery of BWSTT, thanks to the
⁎ Corresponding author at: IRCCS Centro Neurolesi “Bonino-Pulejo”, S.S. 113, Contrada
Casazza, 98124, Messina, Italy.
E-mail address: salbro77@tiscali.it (R.S. Calabrò).
1
These authors equally contributed to the work.
motorized, robotically driven gait orthoses, thus improving over ground
walking ability in individuals with stroke and incomplete spinal cord in-
jury [9–11]. Even though RAGT devices have some limitations, including
clinical and economic issues, they have many advantages over conven-
tional BWSTT methods, such as less effort for the physiotherapist, longer
session duration, higher number of repetitions of task-oriented exer-
cises, physiological and reproducible gait patterns, and the possibility
of measuring patient performance [10].
Virtual Reality (VR) represents an acceptable approximation of the
real world (e.g., walking on an uneven or slippery surface, in a crowded
area, etc.), providing instructive, stimulating, interactive, and direct
feedbacks to enhance the patient's motivation. VR has been successfully
applied to Lokomat (namely Lokomat-Pro) [9,10] to enhance gait per-
formance. Differently from Lokomat-Nanos, Lokomat-Pro is equipped
with an Augmented Performance Feedback, giving patients the oppor-
tunity to playfully exercise functional movements by measuring their
performance and presenting it within task-specific exercises. Such exer-
cises, performed by the patient's avatar in different virtual environment,
can be adapted to the motor and cognitive skills of the patient by
adjusting the intensity and the level of difficulty, thus getting personal-
ized feedback [10,11].
Emerging data in post-stroke individuals and in patients with
Parkinson's disease are suggesting the usefulness of VR in boosting

http://dx.doi.org/10.1016/j.jns.2017.03.047
0022-510X/© 2017 Elsevier B.V. All rights reserved.

Descargado para valentina franco espejo (vfrancoespejo@gmail.com) en Universidad de la Sabana de ClinicalKey.es por Elsevier en septiembre 30, 2018.
Para uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2018. Elsevier Inc. Todos los derechos reservados.
26 R.S. Calabrò et al. / Journal of the Neurological Sciences 377 (2017) 25–30

BWSTT and Lokomat effects on gait ability [12–14], aside from motor
performance and psychological well-being [14]. There are also some
positive data on the usefulness of Lokomat-based training as compared
to conventional gait training and BWSTT [15–20] and of VR in improv-
ing gait and balance in MS patients, but there is no convincing evidence
of the utility of a combined approach employing VR and RAGT in MS
[21–22].
Therefore, the aim of our study was to evaluate the efficacy of
Lokomat-Pro (i.e., RAGT + VR) as compared to Lokomat-Nanos (i.e.,
RAGT-VR) in improving physical and psychological status in patients
with MS.
2. Materials and methods
The present study is a single-blind randomized clinical trial (Clinical
trials ID: NCT02834533), carried at the Robotic Neurorehabilitation Lab-
oratory of the IRCCS Neurolesi Bonino-Pulejo (Messina, Italy). The study
was conducted according to the Declaration of Helsinki, the guidelines
for Good Clinical Practice, and the (CONSORT) Statement guidelines.
The study protocol was approved by our Institutional Review Board
and Ethics Committee (project number: 24/2013).
One hundred and fifty outpatients with relapsing-remitting MS (ac-
cording to Polman criteria) [23] and complaining of gait and/or balance
problems were consecutively screened for study eligibility from January
2015 to January 2016.
Inclusion criteria were: age 18–65 years; moderate to severe walk-
ing disability with Expanded Disability Status Score between 4.0 and
5.5 (pyramidal sub-item ≥ 3) [24]; Montreal Cognitive Assessment
score ≥ 24; absence of concomitant neurological or orthopedic condi-
tions that may interfere with ambulation [25]; stable pharmacological
therapy for at least 6 months. Exclusion criteria were: MS relapse during
the three months prior to recruitment; presence of paroxysmal vertigo;
lower limb botulinum toxin injections within the previous 12 weeks;
cardiorespiratory instability; high-risk of spontaneous fracture
(assessed by Computerized Bone Mineralometry score); lower-limb
skin lesions and phlebitis/thrombosis; N130 kg body weight; visual acu-
ity and visual perception impairment.
According to these inclusion/exclusion criteria, 40 out of 150 pa-
tients were enrolled, randomized (using a simple randomization
scheme generated by a software; www.randomization.com) and allo-
cated into two equally numbered treatment groups: Lokomat-Nanos
(RAGT-VR) or Lokomat-Pro (RAGT + VR) (Fig. 1). Concerning random-
ization, individual, sequentially numbered index cards with the random
assignments were prepared. The index cards were folded and placed in
sealed opaque envelopes. A physician member of the research team,
who was blinded to the baseline examination findings, opened the en-
velopes to attribute the interventions according to the group assign-
ments. All the patients underwent a standard physical treatment
program, consisting of general conditioning exercises (5 min of
warming up, e.g. calf, shoulder, and hand passive range of motion exer-
cises; 5 min of lower and upper extremity strengthening; 20 min of pos-
tural control exercise with maintenance of standing and shifting the
weight loads to the paretic side). After 15 min of rest, both the study
groups received 40 min of RAGT by means of Lokomat (Hocoma Inc.,
Volketswil, Switzerland). The whole protocol was performed five days
a week for eight consecutive weeks.
The Lokomat consists of a treadmill and an exoskeleton with two ac-
tuated orthoses, attached to the participant's limbs with cuffs and
straps. The hip and knee joints of the Lokomat are actuated by linear
drives that move the orthoses through the gait cycle, in the sagittal
plane [26]. For treatment with the robotic system, the amount of body
weight support was initially set at 70% of every patient's weight, then
decreasing in accordance with load tolerance, although not providing
b20% support. The selected speed was adapted to the patient's walking
comfort under the supervision of a trained physiotherapist. Knee flexion
during stance phase of walking was used as an indication for increasing
weight support, and toe off during stance phase as an indication for re-
ducing the weight support. Prior to testing, hip width, length of lower

Fig. 1. Shows the CONSORT flow diagram.

Descargado para valentina franco espejo (vfrancoespejo@gmail.com) en Universidad de la Sabana de ClinicalKey.es por Elsevier en septiembre 30, 2018.
Para uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2018. Elsevier Inc. Todos los derechos reservados.
 R.S. Calabrò et al. / Journal of the Neurological Sciences 377 (2017) 25–30
limbs of the exoskeleton, and size and position of leg cuffs were fitted to
match the anatomy of the participant. Lokomat gait parameters (e.g.
step length, patient coefficient, knee- and hip angles and their respec-
tive offsets) were set so that walking in the device was as natural and
comfortable as possible. All the patients were naïve to robotic assisted
gait training. The accommodation of the patient to Lokomat lasted no
N10 min. The whole RAGT was supervised by a RAGT-trained physio-
therapist, who instructed the patient at the beginning of RAGT on walk-
ing modalities, and checked patient performance and cooperation,
without giving any cue concerning motor performance to avoid extrin-
sic feedbacks. During RAGT, successful and unsuccessful passes, as de-
termined by the inertial measurements, were rendered to the subject
during the trial with a visual feedback (a smiling face), whereas
Lokomat-Pro used an Augmented Feedback Module projecting patient's
avatar while walking on a screen (Fig. 2). About that, subjects were re-
quired to pass obstacles or catch objects appearing on the trail, thus
being forced to change walking direction, and this was achieved by
changing the force exerted by a lower limb as compared to the other.
Hence, patients were aware of their performance and results by observ-
ing their avatar walking on the screen. Specific simulations were chosen
to address the most common multiple sclerosis gait problems (i.e., de-
creased foot clearance, obstacle avoidance, and problems with plan-
ning). Several dynamic distracters were also present in the virtual
environment to challenge subject's attention.
Outcome measures were evaluated at baseline (T0) and after eight
weeks of training (T1) by skilled neurologist blinded on patient alloca-
tion. The primary outcomes were the time up and go test (TUG), the
Berg Balance Scale (BBS), and the Coping Orientation to Problem Expe-
rienced (COPE); the secondary outcomes measures were the Functional
Independence Measure (FIM), Modified Ashworth Scale (MAS), and
Hamilton Rating Scale for Depression (HRSD) (Table 1) [27–31]. More-
over, hip and knee flexion/extension force were evaluated from both
lower limbs by using the Lokomat-Pro device, which assesses the mus-
cle isometric force generated by vastus lateralis, biceps femori, tibialis
anterior, and triceps surae while the patient is in a static position. All
the tests were performed at the beginning (T0) and at the end (T1) of
the rehabilitative program.
The experimenters who analyses the data were different from those
who performed the experiment, and were blind on patient allocation.
Fig. 2. Shows the patient's avatar in different virtual environment with several dynamic targets and/or distracters, i.e. subjects were required to pass obstacles or catch objects appearing on
the trail.
Descargado para valentina franco espejo (vfrancoespejo@gmail.com) en Universidad de la Sabana de ClinicalKey.es por Elsevier en septiembre 30, 2018.
Para uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2018. Elsevier Inc. Todos los derechos reservados.
27
Given the non-normal data distribution (as assessed by the Kolmo-
gorov–Smirnov test), a non-parametric ANOVA for repeated measures
was used to assess the effect of time (two levels: T0 and T1) by group in-
teraction (two levels: RAGT + VR and RAGT-VR) on each parameter
[32]. Tukey's range test was used on comparisons, if the ANOVA was sig-
nificant. For all tests, significance was set at a p-value b 0.05 (two-
tailed). We also calculated the effect size according to Cohen's d stan-
dards for between-group and within-group T0-T1 differences, as to ob-
tain information on the strength of the effect [33,34]. In keeping with
Cohen's standards, effect size results were interpreted as follows: 0.2
is indicative of small, 0.5 moderate, and 0.8 large effect size. Data are re-
ported as median and range [35].
3. Results
There were no significant differences at baseline between the two
groups regarding demographic characteristics. The two main drugs ad-
ministered in all the patients were natalizumab and fingolimod. Both
the groups showed similar EDSS, FSS, and FIM scoring, besides moderate
spasticity, balance and gait abnormalities, and mild depression (Table
2). Also, both samples presented an overall moderate score in nearly
all the COPE sub-items.
Both the RAGT programs were well tolerated, and no patients with-
drew because of either factors related to the RAGT or any adverse effect.
Overall, the effect sizes of differences between RAGT + VR and RAGT-VR
showed a very small effect (b 0.2) for BBS and TUG, even though the pre-
post within-group analysis revealed a moderate effect (~0.6), favoring
RAGT + VR. A very small effect was also reported for the social support,
avoidance, and orientation sub-items of COPE, whereas a large effect
was found for positive attitude (~0.9) and a moderate for problem solv-
ing (~0.5), favoring RAGT + VR (Table 3). In fact, within-group differ-
ences were significant only in the RAGT + VR group, supporting the
findings in the between-group comparison (Table 3).
Our data concerning BBS and TUG are in line with the minimal clin-
ically important difference reported in the literature for people with MS,
i.e., an increase of at least 6 points and a reduction of at least 20%, re-
spectively [36–37].
Given that there are no COPE normative data on minimal detectable
change and minimally clinically important difference in patients with
28 R.S. Calabrò et al. / Journal of the Neurological Sciences 377 (2017) 25–30
Table 1
The main clinical outcome measures.
Test Timed Up and Go (TUG)
This is a simple test used to assess a person's mobility and requires both static and
dynamic balance. It uses the time that a person takes to rise from a chair, walk
three meters, turn around, walk back to the chair, and sit down. During the test,
the person is expected to wear their regular footwear and use any mobility aids
that they would normally require. Scores of ten seconds or less indicate normal
mobility, 11–20 s are within normal limits for disabled patients, and N20 s means
the person needs assistance outside and indicates further examination and
intervention. A score of 30 s or more suggests that the person may be prone to
falls
Berg Balance Scale (BBS)
This is a widely used clinical test of a person's static and dynamic balance abilities.
For functional balance tests, the BBS is generally considered to be the gold
standard. The test takes 15–20 min and comprises a set of 14 simple balance
related tasks, ranging from standing up from a sitting position, to standing on
one foot. The degree of success in achieving each task is given a score of zero
(unable) to four (independent), and the final measure is the sum of all of the
scores. The interpretation of the result is: ≤20 wheelchair user; N20 ≤ 40 walking
with assistance; N40 ≤ 56 independent. Alternatively, the BBS can be used as a
multilevel tool, with the risk of multiple falls increasing below a score of 45 and a
significant increase below 40
Modified Ashworth Scale (MAS)
The Modified Ashworth Scale is considered the primary clinical measure of muscle
spasticity in patients with neurological conditions. It is a five-point scale, with a
grade score of 0, 1, 2, 3, or 4. In 1987, Bohannon and Smith added the grade ‘1+’
and proposed slight changes in the definitions of each score to increase the
sensitivity of the measure and facilitate scoring. The new measure was then
called the Modified Ashworth Scale, and it is considered by many as the gold
standard for measuring spasticity
Expanded Disability Severity Scale (EDSS)
The EDSS provides a total score on a scale that ranges from 0 to 10. The first levels
1.0 to 4.5 refer to people with a high degree of ambulatory ability and the
subsequent levels 5.0 to 9.5 refer to the loss of ambulatory ability. The range of
main categories include (0) = normal neurologic exam; to (5) = ambulatory
without aid or rest for 200 m; disability severe enough to impair full daily activities;
to (10) = death due to MS. In addition, it also provides eight subscale
measurements called Functional System (FS) scores
Coping Orientation to Problem Experienced (COPE)
The COPE is an instrument that measures different coping strategies, generally
used during stressful situations, to assess the different ways in which individuals
respond to stress. It consists of 60 items related to different subscales. Five
subscales (of four items each) measure conceptually distinct aspects of
problem-focused coping (active coping, planning, suppression of competing
activities, restraint coping, seeking of instrumental social support); five
subscales measure aspects of what might be viewed as emotion-focused coping
(seeking of emotional social support, positive reinterpretation, acceptance,
denial, turning to religion); three subscales measure coping responses that are
arguably less useful (focus on and venting of emotions, behavioral
disengagement, mental disengagement)
Hamilton Rating Scale for Depression (HRSD)
The HRS-D is a 17–21-item scale measuring the severity of depressive and
somatization symptoms, where a score of ≥ 15 is generally considered to be
indicative of a diagnosis of depression. The HDRS was originally developed for
hospital in-patients, thus the emphasis on melancholic and physical symptoms
of depression. A later 21-item version (HDRS21) included four items intended to
subtype the depression, but that are sometimes, incorrectly, used to rate
severity. A limitation of the HDRS is that atypical symptoms of depression (e.g.
hypersomnia, hyperphagia) are not assessed
Functional Independence Measure (FIM)
The FIM scale assesses physical and cognitive disability. The scale includes 18 items
with two subscales: motor and socio cognitive. The motor subscale includes 13
items: eating, grooming bathing, dressing upper extremity, dressing lower
extremity, bowel management, bladder management, transfers to bed, chair or
wheelchair, transfer to tub, toilet and shower, walking or wheelchair propulsion,
and stair climbing, whereas the socio cognitive one includes 5 items:
comprehension, expression, social interaction, problem solving, and memory.
Each item is scored from 1 to 7; a score of 1 represents total dependence and a
score of 7 indicates complete independence. The scale can be administered by a
physician, nurse, therapist, or lay person. Possible scores range from 18 to 126,
with higher scores indicating more independence
MS, we have considered a change of at least 20% as clinically relevant in
our sample.
With regard to secondary outcomes (Table 4), mood, hip and knee
strength improved in both the groups at T1, without any significant
Descargado para valentina franco espejo (vfrancoespejo@gmail.com) en Universidad de la Sabana de ClinicalKey.es por Elsevier en septiembre 30, 2018.
Para uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2018. Elsevier Inc. Todos los derechos reservados.
Table 2
Demographic characteristics (median, range).
RAGT-VR (n = 20) RAGT + VR (n = 20)
Age (years) 41(38–47) 44(40–48)
Gender (M/F) 8M/12F 7M/13F
Disease duration (years) 11.5(8–14) 11.5(8–16)
EDSS 4.75(4.1–5.5) 4.4 (4–4.9)
years of education (years) 10(7–13) 11(7–15)
between-group difference (Table 3). MAS and FIM showed no signifi-
cant changes.
4. Discussion and conclusion
There is no convincing evidence that RAGT is superior to conven-
tional walking training or BWSTT in improving gait speed and walking
[12–15]. Nonetheless, it has been shown that RAGT has several advan-
tages as compared to traditional over-ground training and BWSTT in
terms of patient safety, reduced fear of falling, number of steps practiced
(intensity of the training), repeatability, and motor paradigm-induced
fatigue [9–10].
With regard to MS, a few studies have evaluated the effect of RAGT
on gait and other functions in comparison to over-ground gait training
or BWSTT in patients with MS, leading to conflicting results [18–19]. In-
deed, some authors have found that RAGT seems to be effective in in-
creasing walking competency and in restoring the kinematic of the hip
and pelvis in MS subjects [18], whilst other authors demonstrated that
it is unlikely that RAGT is better than over-ground walking training in
patients with an EDSS between 3.0 and 6.5 [19]. We have found that pa-
tients undergoing RAGT have good functional outcomes, but those
performing VR had better results. In fact, nearly all the RAGT + VR pa-
tients showed greater improvement in the primary outcomes (BBS
and COPE), as compared with RAGT-VR individuals. To the best of our
knowledge, this is the first study comparing RAGT with and without
VR in MS patients, whilst a previous study on gait impairment treatment
dealt with intensive and progressive VR coupled with BWSTT [22].
Moreover, we noted a greater improvement in balance in individuals
undergoing VR, in agreement with a previous study [21], further
supporting the idea of the VR pivotal role in motor rehabilitation of
MS patients.
Another relevant finding in our study is that VR can also improve the
psychological outcomes, as shown by coping strategies improvement.
Specifically, the patients using VR got a greater positive attitude and
solving ability toward their clinical problems.
Several issues could explain the fact that patients undergoing
Lokomat-Pro training presented a more evident functional improve-
ment as compared to Lokomat-Nanos.
Indeed, VR represents a valid tool to augment the repetitive practice,
the motivation to endure practice, to promote visual, auditory and tac-
tile input, and motor learning, and to strengthen feedback about perfor-
mance (i.e. knowledge of performance) [11]. The latter is central to
motor learning, since it may induce deep cortical and subcortical chang-
es at the cellular and synaptic level [39], in association with the propri-
oceptive and exteroceptive feedback related to the execution of skilled
tasks. Further, VR provides immediate feedback to performance, thus
assisting with the learning of new motor strategies of movement [21].
Moreover, VR is thought to entrain mirror neuron system thanks to
the visuo-motor information coming from the observation of the
human avatar walking on the screen [40]. Such information recall stored
motor plans, thus contributing to potentiate the motor performance.
Additionally, the significant contribution of VR to RAGT effects may de-
pend on the improvement in either attention/motivation (as the inte-
grated biofeedback system monitors the patient's gait and provides
real-time visual performance feedback to stimulate the patient's active
 R.S. Calabrò et al. / Journal of the Neurological Sciences 377 (2017) 25–30 29

Table 3
shows primary outcome measures at baseline and after 8-week treatment (data are reported as median and range). The differences of change during treatment at within-group and be-
tween-group differences are also reported as effect sizes (95%CI) and p-value.

Parameter Group T0 T1 p-Value, effect size within-group difference (95% CI) p-Value, effect size between-group difference (95% CI)

BBS G1 36 (29;42) 44 (37;51) p = 0.003 p = 0.8

0.255 (−3.128 to 2.617) −0.019 (−2.403 to 2.365)
G2 35 (28;40) 50 (41;58) p b 0.001
0.617 (−3.522 to 2.287)
TUG G1 9.8 (0;15) 8 (2;15) p = 0.002 p = 0.3
−0.243 (−0.36 to 0.847) −0.064 (−0.408 to 0.536)
G2 10 (0;14) 7.9 (2;15) p = 0.001
−0.602 (−0.002 to 1.206)

COPE
Social support G1 27(16;29) 29(19;35) p = 0.03 p = 0.5
0.165 (−1.977 to 1.647) −0.022(−0.691 to 0.645)
G2 22(16;30) 26(19;32) p = 0.02
0.342 (−2.01 to 1.325)
Avoidance G1 23(18;29) 22(17;32) p = 0.01 p = 0.9
−0.1 (−1.41 to 1.61) −0.007 (−0.207 to 0.221)
G2 21(18;24) 20(19;25) p = 0.007
−0.152 (−2.121 to 1.816)
Positive attitude G1 28(22;31) 30(26;35) p = 0.4 p = 0.005
0.225 (−2.194 to 1.743) −0.505 (−3.615 to 2.604)
G2 28(16;37) 33(16;39) p = 0.02
0.225 (−2.194 to 1.743)
Problem-solving G1 27(22;29) 30(26;33) 0.3 p = 0.002
0.243 (−2.056 to 1.568) −0.905 (−2.113 to 0.302)
G2 25(18;31) 32(17;39) p = 0.008
0.243 (−2.056 to 1.568)
Orientation G1 23(10;28) 24(20;27) p = 0.04 p = 0.8
0.1 (−1.61 to 1.41) −0.045 (−0.288 to 0.379)
G2 24(10;28) 22(19;27) p = 0.05
−0.197 (−1.313 to 1.707)

Legend: BBS Berg Balance Scale, COPE Coping Orientation to Problem Experienced, G1 RAGT-VR, G2 RAGT + VR, TUG time up and go test (measured in seconds).

participation) or mood (as it is well known that depression may worsen
cognitive impairment) [14,21].
If the patients are motivated by experiencing a varied and stimulat-
ing environment through the VR, they would get an improvement in at-
tention with potentially better functional outcomes, maybe thanks to
the reactivation/boosting of brain neurotransmission, including the
cholinergic and dopaminergic system [39,40]. In our study, the motiva-
tion necessary to tolerate the extensive practice period was ensured by
providing challenging tasks, including passing obstacles or catching ob-
jects appearing on the trail. Moreover, several dynamic distracters were
added to increase patient's attention.
Even though promising, our results ought to be regarded as prelim-
inary due to some limitations. In particular, we enrolled a small number
of patients, as post-hoc power calculation showed that 120 patients per
group were needed to detect a 20% difference between the RAGT + VR
and RAGT-VR groups (assuming a standard deviation = 6%) with 80%
power (details were inferred from previous studies delivering RAGT in
MS patients) [40–41]. In addition, we have to acknowledge the short
follow-up period and the lack of a conventional over-ground or
BWSTT control group.
In conclusion, our preliminary results indicate that VR may be a valu-
able tool in further improving motor function and psychological well-

Table 4
shows secondary outcome measures at baseline and after 8-week treatment (data are reported as median and range). The differences of change during treatment at within-group and
between-group differences are also reported as effect sizes (95%CI) and p-value.

Parameter Group T0 T1 p-Value, effect size within-group difference (95% CI) p-Value, effect size between-group difference (95% CI)

HRSD G1 12 (3; 27) 7 (2; 21) p = 0.003 p = 0.9

−0.702 (−0.067 to 1.472) −0.062 (−4.932 to 4.808)
G2 10 (3; 24) 6 (3; 24) p b 0.001
−0.668 (−0.006 to 1.344)
HIP G1 37 (11; 118) 49 (15; 89) p = 0.8 p = 0.09
0.913 (0.846 to 0.981) −0.011 (−0.554 to 0.532)
G2 36 (1; 118) 43 (1; 99) p = 0.6
0.846 (0.751 to 0.942)
KNEE G1 30 (0; 59) 36 (1; 77) p b 0.001 p = 0.9
0.372 (−2.642 to 1.897) −0.357 (−8.359 to 9.075)
G2 24 (2; 46) 34 (6; 54) p = 0.006
0.601 (−2.616 to 1.413)
MAS G1 1.5 (0; 2.5) 0.5 (0; 3) p = 0.2 p = 0.4
−0.045 (−3.276 to 3.368) −0.01 (−0.539 to 0.539)
G2 2 (0; 4) 1 (0; 2.5) p = 0.1
−0.206 (−3.08 to 2.666)
FIM G1 89 (80; 95) 92 (88; 101) p = 0.3 p = 0.5
0.073 (−6.267 to 6.12) −0.054 (−2.73 to 2.839)
G2 87 (80; 100) 89 (80; 95) p = 0.4
0.051 (−5.943 to 5.84)

Legend: FIM Functional Independence Measure, G1 RAGT-VR, G2 RAGT + VR, HRSD Hamilton Rating Scale for Depression, MAS Modified Ashworth Scale, hip and knee flexion/extension
force are measured in Newton.

Descargado para valentina franco espejo (vfrancoespejo@gmail.com) en Universidad de la Sabana de ClinicalKey.es por Elsevier en septiembre 30, 2018.
Para uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2018. Elsevier Inc. Todos los derechos reservados.
30 R.S. Calabrò et al. / Journal of the Neurological Sciences 377 (2017) 25–30
being in patients affected by MS. However, further larger sample studies
are needed to investigate the real impact and long term follow-up of this
promising treatment option in MS patients.
Conflict of interest
The authors declare neither conflict of interest nor financial support.
Acknowledgement
The authors wish to thank Mrs. Antonina Donato for English lan-
guage editing.
References
[1] J. Howard, S. Trevick, D.S. Younger, Epidemiology of multiple sclerosis, Neurol. Clin.
34 (4) (2016) 919–939.
[2] M.G. Benedetti, R. Piperno, L. Simoncini, P. Bonato, A. Tonini, S. Giannini, Gait abnor-
malities in minimally impaired multiple sclerosis patients, Mult. Scler. 5 (1999)
363–368.
[3] N. Voloshyna, E. Havrdová, M. Hutchinson, T. Nehrych, X. You, S. Belachew, C.
Hotermans, D. Paes, Natalizumab improves ambulation in relapsing-remitting mul-
tiple sclerosis: results from the prospective TIMER study and a retrospective analy-
sis of AFFIRM, Eur. J. Neurol. 22 (2015) 570–577.
[4] M.H. Rabadi, K. Kreymborg, A.S. Vincent, Sustained-release fampridine (4-
aminopyridine) in multiple sclerosis: efficacy and impact on motor function,
Drugs R D 13 (2013) 175–181.
[5] E. Campbell, E.H. Coulter, P.G. Mattison, L. Miller, A. McFadyen, L. Paul, Physiothera-
py rehabilitation for people with progressive multiple sclerosis: a systematic review,
Arch. Phys. Med. Rehabil. 97 (2016) 141–151.
[6] A. Kalron, D. Nitzani, D. Magalashvili, et al., A personalized, intense physical rehabil-
itation program improves walking in people with multiple sclerosis presenting with
different levels of disability: a retrospective cohort, BMC Neurol. 4 (2015) 15–21.
[7] H. Barbeau, M. Visintin, Optimal outcomes obtained with body-weight support
combined with treadmill training in stroke patients, Arch. Phys. Med. Rehabil. 84
(2003) 1458–1465.
[8] I. Schwartz, A. Sajin, I. Fisher, et al., The effectiveness of locomotor therapy using ro-
botic-assisted gait training in subacute stroke patients: a randomized controlled
trial, PMR 1 (2009) 516–523.
[9] J. Mehrholz, B. Elsner, C. Werner, J. Kugler, M. Pohl, Electromechanical-assisted train-
ing for walking after stroke: updated evidence, Stroke 44 (2013) 127–128 (Review).
[10] R.S. Calabrò, A. Cacciola, F. Bertè, A. Manuli, A. Leo, A. Bramanti, A. Naro, D. Milardi, P.
Bramanti, Robotic gait rehabilitation and substitution devices in neurological disor-
ders: where are we now? Neurol. Sci. 37 (2016) 503–514.
[11] A.S. Merians, D. Jack, R. Boian, et al., Virtual reality-augmented rehabilitation for pa-
tients following stroke, Phys. Ther. 82 (2002) 898–915.
[12] A. Furnari, R.S. Calabrò, M.C. De Cola, M. Bartolo, A. Castelli, A. Mapelli, G. Buttacchio,
E. Farini, P. Bramanti, R. Casale, Robotic-assisted gait training in Parkinson's disease:
a 3-month follow-up randomized clinical trial, Int. J. Neurosci. (2017)http://dx.doi.
org/10.1080/00207454.2017.1288623.
[13] M.L. Walker, S.I. Ringleb, G.C. Maihafer, et al., Virtual reality-enhanced partial body
weight-supported treadmill training post stroke: feasibility and effectiveness in 6
subjects, Arch. Phys. Med. Rehabil. 91 (2010) 115–122.
[14] R.S. Calabrò, M.C. De Cola, A. Leo, et al., Robotic neurorehabilitation in patients with
chronic stroke: psychological well-being beyond motor improvement, Int. J. Rehabil.
Res. 38 (2015) 219–225.
[15] S. Straudi, C. Fanciullacci, C. Martinuzzi, C. Pavarelli, B. Rossi, C. Chisari, N. Basaglia,
The effects of robot-assisted gait training in progressive multiple sclerosis: a ran-
domized controlled trial, Mult. Scler. 22 (2016) 373–384.
[16] M. Gandolfi, C. Geroin, A. Picelli, et al., Robot-assisted vs. sensory integration training
in treating gait and balance dysfunctions in patients with multiple sclerosis: a ran-
domized controlled trial, Front. Hum. Neurosci. 8 (2014) 318.
Descargado para valentina franco espejo (vfrancoespejo@gmail.com) en Universidad de la Sabana de ClinicalKey.es por Elsevier en septiembre 30, 2018.
Para uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2018. Elsevier Inc. Todos los derechos reservados.
[17] S. Beer, B. Aschbacher, D. Manoglou, E. Gamper, J. Kool, J. Kesselring, Robot assisted
gait training in multiple sclerosis: a pilot randomized trial, Mult. Scler. 14 (2008)
231–236.
[18] S. Straudi, M.G. Benedetti, E. Venturini, M. Manca, C. Foti, N. Basaglia, Does robot-
assisted gait training ameliorate gait abnormalities in multiple sclerosis? A pilot ran-
domized-control trial, NeuroRehabilitation 33 (2013) 555–563.
[19] C. Vaney, B. Gattlen, V. Lugon-Moulin, et al., Robotic-assisted step training (lokomat)
not superior to equal intensity of overground rehabilitation in patients with multi-
ple sclerosis, Neurorehabil. Neural Repair 26 (2012) 212–221.
[20] A.C. Lo, E.W. Triche, Improving gait in multiple sclerosis using robotassisted, body
weight supported treadmill training, Neurorehabil. Neural Repair 22 (2008)
661–671.
[21] A. Kalron, I. Fonkatz, L. Frid, H. Baransi, A. Achiron, The effect of balance training on
postural control in people with multiple sclerosis using the CAREN virtual reality
system: a pilot randomized controlled trial, J. Neuroeng. Rehabil. 13 (2016) 13.
[22] A. Peruzzi, A. Cereatti, U. Della Croce, A. Mirelman, Effects of a virtual reality and
treadmill training on gait of subjects with multiple sclerosis: a pilot study, Mult.
Scler. Relat. Disord. 5 (2016) 91–96.
[23] C.H. Polman, S.C. Reingold, B. Banwell, M. Clanet, J.A. Cohen, M. Filippi, K. Fujihara, E.
Havrdova, M. Hutchinson, L. Kappos, F.D. Lublin, X. Montalban, P. O'Connor, M.
Sandberg-Wollheim, A.J. Thompson, E. Waubant, B. Weinshenker, J.S. Wolinsky, Di-
agnostic criteria for multiple sclerosis: 2010 revisions to the McDonald criteria, Ann.
Neurol. 69 (2) (2011 Feb) 292–302.
[24] J.F. Kurtzke, Rating neurologic impairment in multiple sclerosis: an expanded dis-
ability status scale (EDSS), Neurology 33 (11) (1983 Nov) 1444–1452 (PubMed
PMID: 6685237).
[25] Z.S. Nasreddine, N.A. Phillips, V. Bédirian, S. Charbonneau, V. Whitehead, I. Collin, J.L.
Cummings, H. Chertkow, The Montreal cognitive assessment, MoCA: a brief screen-
ing tool for mild cognitive impairment, J. Am. Geriatr. Soc. 53 (4) (2005 Apr)
695–699.
[26] R. Riener, L. Lünenburger, I.C. Maier, G. Colombo, V. Dietz, Locomotor training in sub-
jects with sensori-motor deficits: an overview of the robotic gait orthosis lokomat, J.
Healthcare Eng. 2 (2010) 197–216.
[27] Adapted from. K. Berg, S.L. Wood-Dauphinee, J.L. Williams, Measuring balance in the
elderly: validation of an instrument, Can. J. Public Health 83 (1992) S7–11.
[28] L.B. Krupp, N.G. La Rocca, J. Muir-Nash, A.D. Steinberg, The fatigue severity scale. Ap-
plication to patients with multiple sclerosis and systemic lupus erythematosus,
Arch. Neurol. 46 (1989) 1121–1123.
[29] R.W. Bohannon, M.B. Smith, Interrater reliability of a modified Ashworth scale of
muscle spasticity, Phys. Ther. 67 (1987) 206–207.
[30] C. Sica, C. Novara, S. Dorz, Sanavio coping orientation to problems experienced
(COPE): Italian translation and adaptation, Ezio Giunti Organizzazioni Speciali 223
(1997) 25–34.
[31] Adapted from. Hedlung, Vieweg, The Hamilton rating scale for depression, J. Oper.
Psychiatry 10 (1979) 149–165.
[32] C. Leys, S. Schumann, A nonparametric method to analyze interactions: the adjusted
rank transform test, J. Exp. Social Psychol. 46 (4) (2010) 684–688.
[33] J. Krauth, Distribution-free Statistics an Application-oriented Approach, Elsevier,
Amsterdam, 1988.
[34] J. Cohen, Statistical Power Analysis for the Behavioral Sciences, Lawrence Earlbaum
Associates, Hillsdale, NJ, 1988.
[35] J. Cohen, A power primer, Psych Bull. 112 (1992) 155–159.
[36] S.E. Lord, D.T. Wade, P.W. Halligan, A comparison of two physiotherapy treatment
approaches to improve walking in multiple sclerosis: a pilot randomized controlled
study, Clin. Rehabil. 12 (1998) 477–486.
[37] Y. Nilsagård, C. Lundholm, E. Denison, L.G. Gunnarsson, Predicting accidental falls in
people with multiple sclerosis—a longitudinal study, Clin. Rehabil. 23 (2009)
259–269.
[39] A. Rizzo, P. Requejo, C.J. Winstein, B. Lange, G. Ragusa, A. Merians, J. Patton, P.
Banerjee, M. Aisen, Virtual reality applications for addressing the needs of those
aging with disability, Stud. Health Technol. Inform. 163 (2011) 510–516.
[40] S. Beer, B. Aschbacher, D. Manoglou, E. Gamper, J. Kool, J. Kesselring, Robot-assisted
gait training in multiple sclerosis: a pilot randomized trial, Mult. Scler. 14 (2008)
231–236.
[41] I. Schwartz, A. Sajin, E. Moreh, I. Fisher, M. Neeb, A. Forest, A. Vaknin-Dembinsky, D.
Karusis, Z. Meiner, Robot-assisted gait training in multiple sclerosis patients: a ran-
domized trial, Mult. Scler. 18 (2012) 881–890.