Restorative Neurology and Neuroscience 35 (2017) 307–317 307

DOI 10.3233/RNN-160706
IOS Press

Transcranial direct current stimulation

combined with visuo-motor training
as treatment for chronic stroke patients
Elena L. Pavlovaa,∗ , Påvel Lindberga,b,c , Amirah Khana , Sigurd Ruschkowskid ,
Michael A. Nitschee,f and Jörgen Borga
a Department of Clinical Sciences Karolinska Institute, Danderyd University Hospital, Stockholm, Sweden
b Centre de Psychiatrie et Neurosciences, INSERM U894, Paris, France
c FR3636 Neurosciences, CNRS, Université Paris Descartes, Sorbonne Paris Cité, Paris, France
d Stockholm County Council Innovation, Danderyd University Hospital, Stockholm, Sweden
e Leibniz Research Centre for Working Environment and Human Factors, Technical University of Dortmund,

Dortmund, Germany
f Department of Neurology, University Medical Hospital Bergmannsheil, Bochum, Germany

Abstract.
Background: Recent studies exploring the combined effect of motor learning and transcranial direct current stimulation
(tDCS) for stroke rehabilitation have shown partially conflicting results.
Objective: To test the efficacy of an optimized hand training approach combined with tDCS in stroke patients.
Methods: In the present pilot study we investigated motor effects of four-week training with a visuomotor grip force tracking
task combined with tDCS in 11 chronic stroke patients. Anodal (0.5 mA) or sham tDCS was applied over the primary motor
cortex of the lesioned side for 20 minutes, twice a day, during training.
Results: No difference between the Active and Sham groups in the total upper extremity (UE) Fugl-Meyer Assessment (FMA)
score was found. The most prominent recovery occurred in the shoulder-elbow FMA sub-score; in this segment a significantly
greater improvement in the Active compared to the Sham group was observed up to two months after the intervention. Mean
hold force during the first treatment session predicted the change in the total UE FMA score after treatment.
Conclusion: Four-week visuo-motor training combined with tDCS showed no difference between the Active and Sham
groups in the total UE FMA score, which may be explained by heterogeneity of the degree of recovery in the Active group.
However, the shoulder-elbow FMA sub-score improved significantly more in the Active compared to the Sham group, which
deserves further study.

Keywords: Transcranial direct current stimulation, visuo-motor training, chronic stroke, hand motor function, Fugl-Meyer
assessment

1. Introduction modulates cortical excitability and activity. A grow-

ing number of studies address tDCS treatment in
Transcranial direct current stimulation (tDCS) is different neurological conditions. Numerous proof-
a non-invasive brain stimulation technique, which of-principle studies have been published in the last
decade. Patients with stroke (Nowak, 2009), chronic
∗ Corresponding author: Elena L. Pavlova, Department of Clin-
pain (Rosen, 2009), Parkinson’s disease (Wu, 2008;
ical Sciences Karolinska Institute, Danderyd University Hospital
SE-18288 Stockholm, Sweden. Tel.: +46 73 505 40 81, Fax: +46 Williams, 2009), depression (Nitsche, 2009; Mur-
8 655 77 54; E-mail: [email protected]. phy, 2009) and several other conditions (Flöel, 2014;

0922-6028/17/$35.00 © 2017 – IOS Press and the authors. All rights reserved
308 E.L. Pavlova et al. / tDCS combined with visuo-motor training

Kuo et al., 2014) might benefit from this intervention. group, compared to 3.85 ± 6.4 (3 out of 7 improved
The majority of studies has explored impairments clinically) in the sham group. In accordance with
after stroke, especially hand motor function, which is these data, Bolognini and co-workers (2011) showed
often compromised (Nowak, 2009). tDCS-mediated that constraint-induced movement therapy combined
improvement in forearm motor function after chronic with bihemispheric tDCS improved FMA scores,
stroke ranges from 10% to 30% in different studies while sham stimulation did not. However, constraint-
(Nowak, 2010; Kang et al., 2016). induced therapy requires some preserved extensor
The physiological mechanisms of tDCS are con- muscle function and is not suitable for severely
sidered to comprise some mechanisms relevant for affected patients. Results related to combined use of
learning. NMDA receptors are critical for excitabil- occupational therapy and tDCS (Lindenberg et al.,
ity changes lasting after stimulation, and the resulting 2010; Nair et al., 2011; Mortensen et al., 2015) or
neuroplasticity of glutamatergic synapses may corre- virtual reality therapy and tDCS (Lee et al., 2014;
spond to neurophysiological correlates of learning, Viana et al., 2014) are more heterogeneous ranging
such as long-term potentiation and long-term depres- from a good improvement compared to sham (Lin-
sion (Liebetanz, 2002; Nitsche 2003a, 2003b, 2004). denberg et al., 2010; Lee et al., 2014) to no additive
Motor learning plays an important role in rehabili- effect of tDCS (Mortensen et al., 2015; Viana et al.,
tation (Kang et al., 2016). One potentially relevant 2014).
strategy to increase the efficacy of rehabilitation Rationally, optimization of the training paradigm
is to combine motor learning and tDCS, because should be one way to maximize treatment effects
tDCS might boost re-learning of motor functions via and to achieve clinically relevant results after com-
synergistic physiological effects. In stroke patients, bined treatment. To develop training paradigms suited
several studies investigated an additive effect of tDCS to improve motor function in patients with severely
when combined with different rehabilitation inter- impaired hand function, new approaches to motor
ventions (constraint-induced, occupational, virtual learning are needed. One new approach is directed
reality, robotic training therapies), but the results of to the improvement of fine force control of the hand,
these studies are not consistent. Some studies (Rocha which is a core element in the ability to manip-
et al., 2015; Lee et al., 2014; Kim et al., 2010; Linden- ulate objects with the hand (i.e., dexterity). Some
berg et al., 2010; Nair et al., 2011; Bolognini et al., previous studies have reported an effect of tDCS
2011; Khedr et al., 2013; Wu et al., 2013) report a on motor learning in different visuo-motor track-
significant difference of motor function improvement ing tasks, which require fine-tuning of force control
between active and sham tDCS combined with train- (Antal et al., 2004; Reis et al., 2009). Recently, a
ing, whereas others (Mortensen et al., 2015; Hesse novel visuo-motor tracking task (Lindberg et al.,
et al., 2011; Viana et al., 2014) found no difference. 2009) was validated in a group of 24 patients
Further, the studies demonstrating an additive effect with different levels of motor impairments in the
of tDCS show considerable variability and not all chronic stage after stroke (Lindberg et al., 2012).
patients reached the level of clinical significance after Notably, also patients with severe impairment could
treatment in these studies. perform the task and improved their tracking per-
Though individual and stimulation parameters, formance during repeated blocks similar to control
like stroke severity, time after stroke or duration of subjects.
stimulation, varied between these studies, part of this In this pilot study we explored the effects of com-
variability might be related to the employed rehabil- bined anodal/sham tDCS of the affected primary
itation interventions. Constraint-induced movement motor cortex and visuo-motor grip force tracking
therapy (CIMT), which is considered a more effective training in chronic stroke patients with a broad range
rehabilitation tool than traditional ones (Steven- of hand function impairments during a four-week
son et al., 2012; Batool et al., 2015), might also treatment. We were particularly interested in visuo-
have advantages in combined therapy with tDCS. In motor learning and in the potential transfer of the
one study by Rocha and co-workers (2015) CIMT effects of this treatment to upper extremity motor
with anodal, cathodal or sham tDCS resulted in an function. Furthermore, all sub-divisions of the upper
improvement in the Fugl-Meyer assessment (FMA) extremity were tested separately to delineate the spe-
score by 11.1 ± 2.7 points (all patients improved cific pattern of recovery, since an improvement not
clinically) in the anodal group, and by 7.2 + 4.5 only of the hand/wrist but also in other sub-divisions
(5 out of 7 improved clinically) in the cathodal of the upper extremity was shown in different
 E.L. Pavlova et al. / tDCS combined with visuo-motor training 309

studies (Kawakami et al., 2016; Susanto et al., 2015). accordance with the inclusion/exclusion criteria of
Knowledge about such recovery mechanisms might the IFCN safety guidelines for TMS (Rossini et al.,
help to optimize training paradigms for future com- 2015). Participants gave written informed consent
bined treatments. We also investigated if initial before participation and were compensated for par-
performance parameters could identify responders to ticipating. The study was approved by the regional
the following intervention. ethical review board of Stockholm and conformed to
the Declaration of Helsinki.
2. Material and methods
2.2. Experimental procedure
2.1. Participants
The clinical examination included UE FMA, Wolf
Eleven patients with chronic paresis of the hand motor function test (WMFT), Box and Blocks test
after first ever stroke were enrolled (mean age: (BBT) and maximal grip force measure. Data were
60.4 + 11.9 years, 6 females) from the register of collected before and after the four-week training pro-
the Department of Rehabilitation Medicine, Dan- gram and again two months later. Stroke location was
deryd University Hospital. Three patients had right assessed by structural magnetic resonance imaging in
and eight patients had left hemiparesis (Table 1). all but one patient (Table 1).
All patients were right-handed. Inclusion criteria Ongoing conventional therapy was continued unin-
were first ever stroke, time from stroke more than terrupted and documented. Patients performed a
six months and an ability to voluntarily produce visuo-motor power grip force-tracking task combined
at least some whole hand grip force according to with (active or sham) tDCS (0.5 mA) daily during
clinical examination. Exclusion criteria were any four weeks. Patients were randomly distributed into
other neurological disorder, ongoing psychiatric dis- Active and Sham groups. The Active group patients
order, multiple stroke locations, a history of epileptic received anodal tDCS to the primary motor cortex of
seizures, cardiac pacemakers or metallic implants in the lesioned hemisphere. Patients were blinded to the
the head and neck, electric devices in the body, inabil- experimental conditions.
ity to follow instructions, use of serotoninergic or During the session, patients were sitting in front of
dopaminergic drugs, drug or alcohol addiction. A a computer screen holding a grip force manipulan-
few patients intermittently used psychotropic drugs dum in the affected hand. Two 20 minutes blocks of
(Table 1). None of the participants used antiepilep- stimulation were separated by 20 minutes rest. Two
tic drugs and none was a smoker. The criteria are in 8 min training sessions with a 4 min interval were

Table 1
Baseline characteristics of the patients
Patient sex age stroke lesion site stroke time FMA Fmax CNS-acting
(years) side type from stroke score kg (%)∗ medicine
(years) (0–66)
Active group
patient 1 F 49 R thalamus hemorrhagic 4 35 4.0 (12.5) duloxetine
patient 2 M 60 R thalamus, ischemic 7 50 21.7 (49.0) none
parietal cortex
patient 3 F 66 R n/a n/a 35 13 3.5 (11.7) none
patient 4 F 63 L pons ischemic 1 29 12.0 (51.1) none
patient 5 M 61 L insula ischemic 2 15 2.8 (5.8) sertraline

Sham group
patient 6 M 62 R basal ganglia hemorrhagic 3 11 6.3 (15.8) pramipexol
patient 7 F 84 R insula, ischemic 1 15 2.0 (10.0) none
basal ganglia
patient 8 M 64 R thalamus hemorrhagic 2 41 9.3 (30) none
patient 9 M 67 R insula hemorrhagic 1 49 14 (27.5) none
patient 10 F 35 L thalamus, ischemic 4 23 3.9 (13) none
temporal cortex
patient 11 F 59 R thalamus hemorrhagic 1 46 12.0 (70.6) none
(∗ ) - maximal grip force in the affected hand and percent from performance in the less affected hand.
310 E.L. Pavlova et al. / tDCS combined with visuo-motor training

incorporated into these stimulation blocks. FMA, performance relative to the same force, since error
WMFT and Box and Blocks tests were repeated next of tracking depends on the target force level (Lind-
day and two months after the intervention. berg, 2012). The 10% MVC force level was used to
evaluate tracking parameters at the same effort level
2.3. Power grip force-tracking task (i.e., relative to MVC). The force rising period (ramp
phase) lasted for 2 s, the hold phase for 4 s, followed
During the task, a target force of two different by instantaneous release of force. Duration of one
levels is presented on the computer screen (Fig. 1). run is about eight minutes. Each run included twenty-
Patients can control the cursor on the screen in real- four ramp-and-hold trials at each force level. Relative
time by pressing a power grip force manipulandum error (N.s) during the ramp phase (the total error (pos-
with the affected hand (www.sensix.fr). Maximal grip itive+negative) normalized to the target force level)
strength was measured using the Jamar isometric and release duration, calculated as the time taken to
dynamometer (Hammer & Lindmark, 2003). Patients reduce the grip force (from 75 to 25% of the target
were instructed to follow the target line with the force) at the end of the hold period, were quantified
cursor as close as possible (Lindberg et al., 2010, trial-by-trial.
2012). The output of the calibrated force transducer
(range: 0–1000 N, sensitivity: 0.1 N, www.sensix.fr) 2.4. Transcranial direct current stimulation
was amplified and then sampled at 1 kHz by a CED
Micro1401 running Spike2 (Cambridge Electronic Bipolar stimulation was delivered by a battery-
Design® ). driven electrical stimulator (Magstim Company,
In total, twelve blocks of the task comprising Whiteland, Dyfed, UK) through a pair of rectangu-
of four ramp-hold-and-release target force Trajecto- lar rubber electrodes (6 × 2.5 cm). Two sessions of
ries were performed (Lindberg et al., 2010, 2012). 0.5 mA (0.028 mA/cm2) tDCS of the primary motor
Force during the hold period of each block was cortex were performed daily (weekdays) during a
either 5 N or 10% of maximal voluntary contrac- four week period. Stimulation sessions continued for
tion (MVC). A low absolute force level (5 N) 20 minutes separated by a 20 minutes interval. Ramp
was used for all patients in order to evaluate the duration was 10 seconds. A comparable stimulation
protocol has been proposed to induce late phase plas-
ticity (Monte-Silva et al., 2013). In sham stimulation
experiments, tDCS was delivered for 30 s. Location of
the primary motor cortex was determined with tran-
scranial magnetic stimulation (TMS). Motor cortex
tDCS was applied with the electrode long axis posi-
tioned in the medio-lateral direction with 45 degrees
angle, to target the motor cortex. In all cases, the
return electrode (5 × 7 cm) was placed above the con-
tralateral orbit. Subjects were seated in a comfortable
chair.

2.5. Transcranial magnetic stimulation

TMS was performed at baseline to test for the

integrity of the corticospinal system, during the inter-
vention it was employed to localize the primary motor
cortex. The hot spot of the First Dorsal Interosseous
(FDI) muscle was defined by a standard ‘figure-of-
eight’ 70-mm coil connected to a Magstim Rapid
stimulator (Magstim Company, Whiteland, Dyfed,
Fig. 1. Performance of the visuo-motor grip force tracking test. UK) as the position which resulted in the consistently
The target force of two different levels is presented on a computer largest motor evoked potential (MEP) amplitudes
screen. Patients can control the cursor on the screen in real-time
by pressing a power grip force manipulandum with the affected
of the FDI of the affected hand. The signals were
hand. amplified and filtered (1.59 Hz to 1 kHz, sampling
 E.L. Pavlova et al. / tDCS combined with visuo-motor training 311

rate of 5 kHz), digitized with a micro 1401 AD equal-sized compartments. Subjects are asked to
converter (Cambridge Electronic Design, Cam- transfer as many as possible blocks from one com-
bridge, UK), and recorded by a computer using partment to the other in one minute.
SIGNAL software (Cambridge Electronic Design,
version 2.13). 2.9. Data analysis

2.6. Fugl-Meyer upper extremity assessment Data were analysed with STATISTICA 12 software
(StatSoft, Inc., USA). Baseline differences between
Fugl-Meyer upper extremity assessment (FMA) the groups in the motor part of the FMA score,
was the primary outcome measure. FMA was devel- WMFT-FAS, age and time since stroke were assessed
oped to evaluate the sensorimotor status after stroke by a one-way ANOVA, and differences in in sex,
(Fugl-Meyer et al., 1975; Gladstone et al., 2002). stroke location – cortico-subcortical or subcortical,
According to the International Classification of Func- and stroke side – by the Chi-square test. Performance
tioning, Disability and Health (ICF) (Sivan et al., in the motor part of the FMA score, all FMA sub-
2011), it relates to body function and structural scores, WMFT-FAS, WMFT-TIME, Box and Blocks
domains (Santisteban et al., 2016). The assessment and grip force tests during the time course of the
includes sections testing shoulder/elbow, wrist, hand, study across the three assessments were explored
coordination/speed, sensation, passive joint move- by mixed-model ANOVAs with TIME as within-
ment and joint pain. The motor part of FMA (total subject factor and GROUP as between-subject factor.
upper extremity (UE) FMA), which covers the first Post-hoc analysis was done with the Fisher LSD
four of the listed sections, has a maximal score of 66, test.
better performance corresponds to higher values. Per- Parameters of the power grip force-tracking task
formance of each task is rated on a 3-point scale (from were analysed with MatLab 2012 (MatLab Inc,
0 to 2). Improvement was classified as clinically sig- USA). Ramp Error, Hold Error, Release Duration and
nificant if an increase in the total UE FMA was Hold Force were calculated for the second sessions
>5 points (Page et al., 2012). FMA was performed of all days. These parameters were grouped into 4
by a physiotherapist blinded to the experimental weeks of training for each patient and were explored
conditions. with mixed model ANOVAs. Post-hoc analysis was
done with the Fisher LSD test.
2.7. Wolf motor function test Simple regression analyses were applied for inves-
tigation of 1) the relation of baseline FMA score,
The Wolf motor function test (WMFT) served as baseline shoulder/elbow sub-score of FMA, time
secondary outcome measure. It evaluates the prox- from stroke, handedness, age and sex and the change
imal and distal motor activity in the UE of adults in FMA score after treatment to identify clinical pre-
with hemiparesis. The test includes timing measures dictors of outcome; 2) the relation between change
(WMFT-TIME) and quality of movement measures in grip force and percent change of grip force and
as assessed by functional ability scores (WMFT- change in FMA scores to delineate if force change
FAS). 15 of the 17 WMFT tasks are timed, 120 predicts clinical outcome; 3) the relation of the Ramp
seconds are allowed for performance of each task. Error, Hold Error, Release Duration and Hold Force
WMFT-FAS is an evaluation based on a 6-point func- during the first training session and change in FMA
tional ability scale (from 0 to 5) with a maximum sum score/shoulder-elbow FMA to reveal early functional
score of 75 (Pereira et al., 2011). predictors.

2.8. Box and Blocks test

3. Results
The Box and Block test (BBT) is routinely
used in clinical practice in order to assess man- Eleven patients participated in the study (6 women,
ual dexterity in patients with decreased hand 5 men, mean age 60.7 SD 11.4). Five patients were
functional abilities. The test is reliable for clin- in the Active group, six in the Sham group. Groups
ical evaluations (Chen et al., 2009) and consists did not differ for demographic characteristics (age,
of blocks (2.5 cm × 2.5 cm × 2.5 cm) in a wooden sex, stroke location – cortico-subcortical or subcor-
box (53.7 cm × 8.5 cm × 27.4 cm) divided into two tical, time since stroke, stroke side – see Table 1).
312 E.L. Pavlova et al. / tDCS combined with visuo-motor training

Fig. 2. (A) Total UE FMA score (0–66) before, immediately after and at 2 months follow-up in the Active and Sham groups. Mixed
model ANOVA with TIME as within-subject factor and GROUP as between-subject factor. TIME: F(2, 18) = 8.05, p = 0.003; GROUP: F(1,
9) = 0.04, p = 0.85; TIME*GROUP: (2, 18) = 0.93, p = 0.41. (B) shoulder/elbow FMA sub-score (0–36) before, immediately after and at 2
months follow-up in the Active and Sham groups. Mixed model ANOVA with TIME as within-subject factor and GROUP as between-subject
factor. TIME: F(2, 18) = 12,931, p = 0.0003; GROUP: (F(1, 9) = 0.00003, p = 1.0; TIME*GROUP: (F(2, 18) = 3.58, p = 0.049). The shoulder-
elbow sub-score increased significantly after the intervention in the Active group (p = 0.00006). Though these elevated values decreased two
months later (p = 0.01), they remained significantly higher two months after the intervention (p = 0.03). (Fisher LSD test).  - mean, bars
–0.95 confidence interval.

All patients were right-handed. No adverse effects Table 2

were observed in any of the participants. Results of the mixed model ANOVAs for movement parameters
Group Time TIME*GROUP

3.1. Group effects Force level: 5N

Ramp Error 0.22 0.61 0.28
Hold Error 0.96 0.17 0.53
No difference in total UE FMA was observed Release duration 0.84 0.67 0.40
at baseline (F(1, 9) = 0.07, p = 0.79). UE FMA Mean hold force 0.71 0.84 0.89
changed significantly over time (factor TIME, F(2,
Force level: 10%
18) = 8.05, p = 0.003, partial eta-squared = 0.47), but Ramp Error 0.68 0.08 0.03
no difference between groups (factor GROUP, F(1, Hold Error 0.36 0.03 0.75
9) = 0.04, p = 0.85) or interaction between these fac- Release duration 0.78 0.88 0.97
Mean hold force 1.00 0.10 0.55
tors (F(2, 18) = 0.93, p = 0.41) was revealed (Fig. 2B).
Bold indicates significant values (p < 0.05).
The shoulder-elbow sub-score of UE FMA sig-
nificantly changed over time (factor TIME, F(2,
18) = 12,931, p = 0.0003, partial eta-squared 0.59) (F(2, 18) = 2.52, p = 0.11 and F(2, 18) = 1.21, p = 0.32,
but showed also a significant interaction between respectively).
TIME and GROUP (F(2, 18) = 3.58, p = 0.049, par- No change was observed in the Box and Blocks test
tial eta-squared 0.28). The shoulder-elbow sub-score (TIME, p = 0.70; GROUP, p = 0.12 GROUP*TIME,
increased significantly after the intervention in the p = 0.40).
Active group (p = 0.00006). Though these elevated A borderline significance was observed for the
values decreased two months later (p = 0.01), they grip MVC change over the time course of the study
were still significantly higher two months after the (TIME, p = 0.05; GROUP, p = 0.88; GROUP*TIME;
intervention than before the intervention (p = 0.03). p = 0.69).
No difference in the factor GROUP was revealed Among the analysed visuo-motor test parameters,
(F(1, 9) = 0.00003, p = 1.0). Sub-scores for the wrist, Hold Error at 10% decreased significantly over the
hand, coordination/speed, sensation, passive joint time course of the study (TIME, p = 0.03; GROUP,
movement and joint pain segments did not change p = 0.36; GROUP*TIME; p = 0.75; Table 2, Fig. 3A).
significantly (not shown). The decrease emerged after one week treatment and
WMFT-FAS and WMFT-TIME did not show any plateaued later (first week of training differed signif-
significant change in either TIME (F(2, 18) = 1.77, icantly from all other weeks, p < 0.05). Ramp Error
p = 0.20 and F(2, 18) = 0.71, p = 0.51, respectively), at 10% decreased only trendwise over time (TIME,
GROUP (F(1, 9) = 0.31, p = 0.59 and F(1, 9) = 0.09, p = 0.08; GROUP, p = 0.68; Table 2). However, a sig-
p = 0.78, respectively) or the respective interaction nificant GROUP*TIME interaction (p = 0.03, partial
 E.L. Pavlova et al. / tDCS combined with visuo-motor training 313

Fig. 3. A. Hold Error during the four weeks of training at 10% MVC force level. Fisher post-hoc tests reveal a significant difference between
first and all other weeks (p < 0.05).  - mean, bars represent 0.95 confidence interval. B. Ramp Error during four weeks of training at 10%
MVC force level. Fisher post-hoc tests reveal a significant difference between first and other weeks in the Active group (p < 0.05).  - mean,
bars represent 0.95 confidence interval.

Fig. 4. Total UE FMA score before, immediately after and 2 months after intervention in individual patients in the Active (A) and Sham (B)
groups. The asterisks indicate clinically significant differences (total UE FMA score >5).

eta-squared 0.27) was observed. Ramp Error in the r2 = 0.63; p = 0.003, r2 = 0.65 and p = 0.23, r2 = 0.16,
Active group significantly decreased after one week respectively).
and plateaued thereafter (Fig. 3B). Mean Hold Force at 10% during the first treat-
ment session predicted the change in the FMA motor
3.2. Predictors of change in hand motor function score after treatment (p = 0.005; R2 = 0.61; Fig. 5)
but not the change of shoulder-elbow FMA (p = 0,14;
No correlations were observed between baseline r2 = 0,23).
FMA score, shoulder/elbow sub-score of FMA, time
from stroke, handedness, age or sex and the change
in FMA score after treatment. Clinically signifi- 4. Discussion
cant improvement according to the FMA score was
revealed in four out of five patients of the Active group The aim of our study was to explore a possible
and in two out of six in the Sham group (Fig. 4). additive effect of tDCS on motor training with a
The most prominent improvement in these patients visuo-motor grip tracking task in patients with motor
was observed in the shoulder/elbow sub-score of the deficits in the chronic stage after stroke. The results
FMA. For example, in patient 1 nine out of ten points indicate that this combined treatment can result in
of the FMA change occurred in this sub-score. clinically significant improvement of upper extrem-
The percent change in grip MVC after treat- ity motor function and also provide information that
ment compared to the before treatment condition may guide further larger studies.
(but not the absolute change in grip MVC) cor- Most improvement occurred in the shoulder-elbow
related with motor FMA and the shoulder-elbow sub-score of FMA. In this segment (but not in
sub-score, but not with the hand sub-score (p = 0.004, the total UE FMA score), a significantly greater
314 E.L. Pavlova et al. / tDCS combined with visuo-motor training

et al., 2015) describing improvement in the shoul-

der/elbow but not in the wrist-hand sub-score of the
FMA after robot assisted finger training. This might
reflect a proximal to distal pattern of recovery after
stroke (Twitchell, 1951) only loosely connected to
the specific treatment applied.
Significant improvement in this domain was main-
tained up to two months after the intervention.

4.2. Learning of the visuo-motor grip force

training task

The visuo-motor grip force tracking task is directed

Fig. 5. Correlation between normalized mean hold force during to the improvement of grip force control in the
the first treatment session at 10% of the maximal grip force and
change in the motor part of the FMA. Simple regression analysis
hand. In our study, a significant decrease of the Hold
(p = 0.005; R2 = 0.61). Error independent from the stimulation protocol was
observed after one week of training. After training,
improvement in the Active compared to the Sham this parameter approached the level of age-matched
group was observed. controls in the previous study (Lindberg et al., 2012).
Furthermore, in all patients Hold error diminished Our study shows that daily grip force-tracking train-
after one week of treatment. Ramp error decreased ing leads to enhanced accuracy of force maintenance.
only in Active group. Notably, treatment results were Ramp error, however, diminished only in the Active
predicted by mean hold force in the first session of group. This might correspond to the additive benefit
the treatment. of tDCS on the accuracy of dynamic force modu-
lation. Although no consistent changes at the group
4.1. Visuo-motor grip force tracking task level were found in either release duration nor mean
combined with tDCS for rehabilitation of hold force.
chronic stroke patients
4.3. Variability of combined treatment effects
Although tDCS showed only a limited additive
benefit to patients in this study and no difference In the Active group, a substantial variability of
between the Active and Sham groups was observed effects was observed (from –3 to 10 points change in
for the primary outcome measure of our study - the FMA). The sham group showed more homogenous
motor part of FMA, the number of patients with clini- results: the FMA change varied from 1 to 5. In our
cally meaningful improvement in the active and sham study a wide range of patients was enrolled: FMA
groups in this pilot study is similar or even exceed scores from 11 to 50, left and right stroke lesions,
those of other studies employing tDCS combined dominant and non-dominant hemiparesis; subcortical
with occupational therapy (Lindenberg et al., 2010; and cortico-subcortical lesions. Therefore, individual
Nair et al., 2011). tDCS combined with more inten- clinical characteristics of the patients could explain
sive constraint-induced movement therapy (6 hours the heterogeneity of the results and, as a result, lack
per day) nevertheless resulted in better outcomes with of whole group differences between active and sham
nearly all patients showing a clinically significant groups in the FMA motor score. Known predictors of
improvement (Rocha et al., 2015). However, a poten- tDCS treatment outcome include severity of stroke
tial advantage of training with the visuo-motor grip (Marquez et al., 2015), level of stoke (cortical/sub-
force tracking task compared to other types of train- cortical) (Bradnam et al., 2012; Marquez et al., 2013;
ing, for example, constraint-induced therapy, is that O’Shea et al., 2014), size of lesion (Bolognini et
it can be used for the whole range of hand dysfunc- al., 2014), integrity of cortical connections (Bradnam
tion, including patients with severe paresis (Lindberg et al., 2012; Lindenberg et al., 2012) and time after
et al., 2012). stroke (Marquez et al., 2013; O’Shea et al., 2014).
Though the greatest change in the shoulder-elbow None of these or other available clinical parameters
sub-score of the FMA was unexpected, these data predicted the effect of the treatment, presumably due
are similar to the results of another study (Susanto to the limited number of patients.
 E.L. Pavlova et al. / tDCS combined with visuo-motor training 315

However, mean hold force in the first session pre- Acknowledgments

dicted long-term effects of the intervention on total
upper extremity motor function (but not on the elbow- The authors are grateful to Jeanette Plantin, Gaia
shoulder part). It is likely that muscular endurance Valentina Pennati, Karin Weber and Negar Chamsai
along with the ability to maintain focused attention for the help with the data collection.
to the visual display contribute to maintaining force
level (Cogiamanian et al., 2007).
Funding

5. Study limitations and future directions The study was funded by Promobilia foundation
(grants D111722003 and D111722023) and Stroke
The small sample size is the main limitation of Riksförbundet (grant D111722033).
the study. In addition, the effect size of the signifi-
cant findings is rather small. Therefore, one should be
cautious about generalization of the results, observed References
in this group, such as predominant improvement of
shoulder-elbow function after visuo-motor training Antal, A., Nitsche, M.A., Kincses, T.Z., Kruse, W., Hoffmann,
K.P., & Paulus, W. (2004). Facilitation of visuo-motor learn-
combined with tDCS. This finding needs to be con- ing by transcranial direct current stimulation of the motor
firmed in larger studies, which should include patients and extrastriate visual areas in humans. European Journal of
with a whole range of upper extremity dysfunction. Neuroscience, 19, 2888-2892.
Different factors, like duration of the disease, can Batool, S., Soomro, N., Amjad, F., & Fauz, R. (2015). To compare
also influence the pattern of recovery after combined the effectiveness of constraint induced movement therapy ver-
sus motor relearning programme to improve motor function
treatment, which should also be investigated in forth- of hemiplegic upper extremity after stroke. Pakistan Journal
coming trials. of Medical Sciences, 31, 1167-1171.
In addition, delineation of the predictors of effects Bolognini, N., Vallar, G., Casati, C., Latif, L.A., El-Nazer, R.,
of combined long-term treatment, based on single Williams, J., Banco, E., & Fregni, F. (2011). Neurophys-
session parameters of motor performance, might be iological and behavioral effects of tDCS combined with
useful. So far, patients with the best chance to benefit constraint-induced movement therapy in poststroke patients.
Neurorehabilitation and Neural Repair, 25, 819-829.
from tDCS according to clinical characteristics are
Bradnam, L.V., Stinear, C.M., Barber, P.A., & Byblow, W.D.
usually enrolled in the tDCS studies. Because of this, (2012). Contralesional hemisphere control of the proximal
only one out of ten available stroke patients is usu- paretic upper limb following stroke. Cerebral Cortex, 22,
ally enrolled (Rocha et al., 2015; Mortensen et al., 2662-2671.
2015; Viana et al., 2014). A motor task-based pre- Chen, H.M., Chen, C.C., Hsueh, I.P., Huang, S.L., & Hsieh, C.L.
dictive approach might be a better approximation of (2009). Test-retest reproducibility and smallest real difference
the effects of combined treatment and cover larger of 5 hand function tests in patients with stroke. Neuroreha-
bilitation and Neural Repair, 23, 435-440.
portions of patients.
Cogiamanian, F., Marceglia, S., Ardolino, G., Barbieri, S., & Priori,
A. (2007). Improved isometric force endurance after transcra-
6. Conclusions nial direct current stimulation over the human motor cortical
areas. European Journal of Neuroscience, 26, 242-249.
Flöel, A. (2014). tDCS-enhanced motor and cognitive function in
Four-week treatment with visuo-motor grip force neurological diseases. Neuroimage, 85(Pt 3), 934-947.
tracking training combined with anodal tDCS in the
Fugl-Meyer, A.R., Jaasko, L., Leyman, I., Olsson, S., & Steglind,
present study resulted in clinical improvement of S. (1975). The post-stroke hemiplegic patient. 1. a method for
motor functions in some patients, but did not show evaluation of physical performance. Scandinavian Journal of
a difference between Active and Sham groups in the Rehabilitation Medicine, 7, 13-31.
primary outcome measure of the study – the motor Gladstone, D.J., Danells, C.J., & Black, S.E. (2002). The
part of FMA, presumably, due to heterogeneity of the Fugl–Meyer assessment of motor recovery after stroke: A
critical review of its measurement properties. Neurorehabili-
degree of recovery. The most prominent improvement tation and Neural Repair, 16, 232-240.
was documented for the shoulder-elbow sub-score of
Hammer, A., & Lindmark, B. (2003). Test-retest intra-rater reli-
FMA, which was significantly larger in the Active ability of grip force in patients with stroke. Journal of
group after intervention. If this observation is con- Rehabilitation Medicine, 35, 189-194.
firmed in larger studies, it might guide training Hesse, S., Waldner, A., Mehrholz, J., Tomelleri, C., Pohl, M.,
protocol optimization in future combined treatments. & Werner, C. (2011). Combined transcranial direct current
316 E.L. Pavlova et al. / tDCS combined with visuo-motor training

stimulation and robot assisted arm training in subacute stroke intracerebral hemorrhage: A double-blind randomized con-
patients: An exploratory, randomized multicenter trial. Neu- trolled trial. Disability and Rehabilitation, 38, 637-643.
rorehabilitation and Neural Repair, 25, 838-846.
Murphy, D.N., Boggio, P., & Fregni, F. (2009). Transcranial direct
Kang, N., Summers, J.J., & Cauraugh, J.H. (2016). Transcranial current stimulation as a therapeutic tool for the treatment
direct current stimulation facilitates motor learning post- of major depression: Insights from past and recent clinical
stroke: A systematic review and meta-analysis. Journal of studies. Current Opinion in Psychiatry, 22, 306-311.
Neurology, Neurosurgery & Psychiatry, 87, 345-355.
Nair, D.G., Renga, V., Lindenberg, R., Zhu, L., & Schlaug, G.
Khedr, E.M., Shawky, O.A., El-Hammady, D.H., Rothwell, J.C., (2011). Optimizing recovery potential through simultane-
Darwish, E.S., Mostafa, O.M., & Tohamy, A.M. (2013). ous occupational therapy and non-invasive brain-stimulation
Effect of anodal versus cathodal transcranial direct cur- using tDCS. Restorative Neurolology and Neuroscience, 29,
rent stimulation on stroke rehabilitation: A pilot randomized 411-420.
controlled trial. Neurorehabilitation and Neural Repair, 27,
Nitsche, M.A., Fricke, K., Henschke, U., Schlitterlau, A., Liebe-
592-601.
tanz, D., Lang, N., . . . , Paulus, W. (2003a). Pharmacological
Kim, D.Y., Lim, J.Y., Kang, E.K., You, D.S., Oh, M.K., Oh, B.M., modulation of cortical excitability shifts induced by tran-
& Paik, N.J. (2010). Effect of transcranial direct current stim- scranial direct current stimulation in humans. Journal of
ulation on motor recovery in patients with subacute stroke. Physiology, 553(Pt 1), 293-301.
American Journal of Physical Medicine & Rehabilitation, 89,
Nitsche, M.A., Schauenburg, A., Lang, N., Liebetanz, D., Exner,
879-886.
C., Paulus, W., & Tergau, F. (2003b). Facilitation of implicit
Kuo, M.F., Paulus, W., & Nitsche, M.A. (2014). Therapeutic effects motor learning by weak transcranial direct current stimula-
of non-invasive brain stimulation with direct currents (tDCS) tion of the primary motor cortex in the human. Journal of
in neuropsychiatric diseases. Neuroimage 85(Pt 3), 948-960. Cognitive Neuroscience, 15, 619-626.
Lee, S.J., & Chun, M.H. (2014). Combination transcranial direct Nitsche, M.A., Jaussi, W., Liebetanz, D., Lang, N., Tergau, F., &
current stimulation and virtual reality therapy for upper Paulus, W. (2004). Consolidation of human motor cortical
extremity training in patients with subacute stroke. Arch Phys neuroplasticity by D-cycloserine. Neuropsychopharmacol-
Medical Rehabilitation, 95, 431-438. ogy, 29, 1573-1578.
Liebetanz, D., Nitsche, M.A., Tergau, F., & Paulus, W. (2002). Nitsche, M.A., Boggio, P.S., Fregni, F., & Pascual-Leone, A.
Pharmacological approach to the mechanisms of transcranial (2009). Treatment of depression with transcranial direct cur-
DC-stimulation-induced after-effects of human motor cortex rent stimulation (tDCS): A review. Expimental Neurology,
excitability. Brain, 125(Pt 10), 2238-2247. 219, 14-19.
Lindberg, P., Ody, C., Feydy, A., & Maier, M.A. (2009). Precision O’Shea, J., Boudrias, M.H., Stagg, C.J., Bachtiar, V., Kischka,
in isometric precision grip force is reduced in middle-aged U., Blicher, J.U., & Johansen-Berg, H. (2014). Predicting
adults. Experimental Brain Research, 193, 213-224. behavioural response to TDCS in chronic motor stroke. Neu-
Lindberg, P.G., Feydy, A., & Maier, M.A. (2010). White mat- roimage, 85(Pt 3), 924-933.
ter organization in cervical spinal cord relates differently to Page, S.J., Fulk, G.D., & Boyne, P. (2012). Clinically important dif-
age and control of grip force in healthy subjects. Journal of ferences for the upperextremity Fugl–Meyer Scale in people
Neuroscience, 30, 4102-4109. with minimal to moderate impairment due to chronic stroke.
Lindberg, P.G., Roche, N., Robertson, J., Roby-Brami, A., Bussel, Physical Therapy, 92, 791-798.
B., & Maier, M.A. (2012). Affected and unaffected quantita- Pereira, N.D., Michaelsen, S.M., Menezes, I.S., Ovando, A.C.,
tive aspects of grip force control in hemiparetic patients after Lima, R.C., & Teixeira-Salmela, L.F. (2011). Reliability of
stroke. Brain Research, 1452, 96-107. the Brazilian version of the Wolf Motor Function Test in adults
Lindenberg, R., Renga, V., Zhu, L.L., Nair, D., & Schlaug, G. with hemiparesis. The Brazilian Journal of Physical Therapy,
(2010). Bihemispheric brain stimulation facilitates motor 15, 257-265.
recovery in chronic stroke patients. Neurology, 75, 2176- Nowak, D.A., Grefkes, C., Ameli, M., & Fink, G.R. (2009). Inter-
2184. hemispheric competition after stroke: Brain stimulation to
Lindenberg, R., Zhu, L.L., Rüber, T., & Schlaug, G. (2012). Pre- enhance recovery of function of the affected hand. Neurore-
dicting functional motor potential in chronic stroke patients habilitation and Neural Repair, 23, 641-656.
using diffusion tensor imaging. Human Brain Mapping, 33, Nowak, D.A., Bösl, K., Podubeckà, J., & Carey, J.R. (2010). Non-
1040-1051. invasive brain stimulation and motor recovery after stroke.
Marquez, J., van Vliet, P., McElduff, P., Lagopoulos, J., & Parsons, Restorative Neurology and Neuroscience, 28, 531-544.
M. (2015). Transcranial direct current stimulation (tDCS): Reis, J., Schambra, H.M., Cohen, L.G., Buch, E.R., Fritsch, B.,
Does it have merit in stroke rehabilitation? A systematic Zarahn, E., & Krakauer, J.W. (2009). Noninvasive cortical
review. International Journal of Stroke, 10, 306-316. stimulation enhances motor skill acquisition over multiple
Monte-Silva, K., Kuo, M.F., Hessenthaler, S., Fresnoza, S., Liebe- days through an effect on consolidation. Proceedings of the
tanz, D., Paulus, W., & Nitsche, M.A. (2013). Induction of National Academy of Sciences, 106, 1590-1595.
late LTP-like plasticity in the human motor cortex by repeated Rocha, S., Silva, E., Foerster, Á., Wiesiolek, C., Chagas, A.P.,
non-invasive brain stimulation. Brain Stimulation, 6, 424-432. Machado, G., & Monte-Silva, K. (2016). The impact of tran-
Mortensen, J., Figlewski, K., & Andersen, H. (2016). Combined scranial direct current stimulation (tDCS) combined with
transcranial direct current stimulation and home-based occu- modified constraint-induced movement therapy (mCIMT)
pational therapy for upper limb motor impairment following on upper limb function in chronic stroke: A double-blind
 E.L. Pavlova et al. / tDCS combined with visuo-motor training 317

randomized controlled trial. Disability and Rehabilitation, 38, Stevenson, T., Thalman, L., Christie, H., & Poluha, W. (2012).
653-660. Constraint-induced movement therapy compared to dose-
matched interventions for upper-limb dysfunction in adult
Rosen, A.C., Ramkumar, M., Nguyen, T., & Hoeft, F. (2009). Non-
invasive transcranial brain stimulation and pain. Current Pain survivors of stroke: A systematic review with meta-analysis.
Physiotherapy Canada, 64, 397-413.
and Headache Reports, 13, 12-17.
Twitchell, T.E. (1951). The restoration of motor function following
Rossini, P.M., Burke, D., Chen, R., Cohen, L.G., Daskalakis, Z.,
hemiplegia in man. Brain, 74, 443-480.
Di Iorio, R., & Ziemann, U. (2015). Non-invasive electrical
and magnetic stimulation of the brain, spinal cord, roots and Viana, R.T., Laurentino, G.E., Souza, R.J., Fonseca, J.B., Silva
peripheral nerves: Basic principles and procedures for routine Filho, E.M., Dias, S.N., . . . , Monte-Silva, K.K. (2014).
clinical and research application. An updated report from an Effects of the addition of transcranial direct current stimula-
I.F.C.N. Committee. Clinical Neurophysiology, 126, 1071- tion to virtual reality therapy after stroke: A pilot randomized
1107. controlled trial. Neuro Rehabilitation, 34, 437-446.
Santisteban, L., Térémetz, M., Bleton, J.P., Baron, J.C., Maier, Williams, J.A., Imamura, M., & Fregni, F. (2009). Updates on the
M.A., & Lindberg, P.G. (2016). Upper limb outcome mea- use of non-invasive brain stimulation in physical and reha-
sures used in stroke rehabilitation studies: A systematic bilitation medicine. Journal of Rehabilitation Medicine 41,
literature review. PLoS One, 11(5), e0154792. 305-311.
Sivan, M., O’Connor, R.J., Makower, S., Levesley, M., & Bhakta, Wu, A.D., Fregni, F., Simon, D., Deblieck, K., & Pascual-Leone,
B. (2011). Systematic review of outcome measures used in C.A. (2008). Noninvasive brain stimulation for Parkinson’s
the evaluation of robot-assisted upper limb exercise in stroke. disease and dystonia. Neurotherapeutics, 5, 345-361.
Journal of Rehabilitation Medicine, 43, 181-189.
Wu, D., Qian, L., Zorowitz, R.D., Zhang, L., Qu, Y., & Yuan, Y.
Susanto, E.A., Tong, R.K., Ockenfeld, C., & Ho, N.S. (2015). (2013). Effects on decreasing upper-limb poststroke muscle
Efficacy of robot-assisted fingers training in chronic stroke tone using transcranial direct current stimulation: A random-
survivors: A pilot randomized-controlled trial. Journal of ized sham-controlled study. Archives of Physical Medicine
Neuroengeneering and Rehabilitation, 25, 12-42. and Rehabilitation, 94, 1-8.