Restorative Neurology and Neuroscience 30 (2012) 497–510 497

DOI 10.3233/RNN-2012-120227
IOS Press

Inter-hemispheric coupling changes associate

with motor improvements after robotic stroke
rehabilitation
G. Pellegrinoa,1,∗ , L. Tomasevicb,1 , M. Tombinia , G. Assenzaa , M. Bravic , S. Sterzic , V. Giacobbed ,
L. Zollod , E. Guglielmellid , G. Cavallod , F. Vernieria and F. Tecchiob
a Department of Neurology, Campus Bio-Medico University, Via A. del Portillo, Rome, Italy
b LET’S – ISTC – CNR – Ospedale Fatebenefratelli – Via di Ponte Quattro Capi, Rome, Italy
c Medicina Fisica e Riabilitazione – Campus Bio-Medico University, Via A. del Portillo, Rome, Italy
d Laboratory of Biomedical Robotics and EMC – Campus Bio-Medico University, Via A. del Portillo, Rome, Italy

Abstract. Purpose: In the chronic phase of stroke brain plasticity plays a crucial role for further motor control improvements.
This study aims to assess the brain plastic reorganizations and their association with clinical progresses induced by a robot-aided
rehabilitation program in chronic stroke patients.
Methods: 7 stroke patients with an upper limb motor impairment in chronic phase underwent a multi-modal evaluation
before starting and at the end of a 12-week upper-limb neurorehabilitation program. Fugl-Meyer Assessment (FMA) Scale
scores and performance indices of hand movement performance (isometric pinch monitored through a visual feedback) were
collected. Cerebral reorganizations were characterized by 32-channel electroencephalography (EEG) focusing on ipsilesional
and contralesional resting state properties investigating both bipolar derivations overlying the middle cerebral artery territory
and the primary somatosensory sources (S1) obtained through the Functional Source Separation (FSS) method. Power Spectral
Density (PSD) and interhemispheric coherence (IHCoh) at rest were measured and correlated with clinical and hand control
robot-induced improvements.
Results: After the robotic rehabilitation we found an improvement of FMAS scores and hand motor control performance and
changes of brain connectivity in high frequency rhythms (24–90 Hz). In particular, the improvement of motor performance
correlated with the modulation of the interhemispheric S1 coherence in the high beta band (24–33 Hz).
Conclusions: Recently it has been shown that an upper limb robot-based rehabilitation improves motor performance in stroke
patients. We confirm this potential and demonstrate that a robot-aided rehabilitation program induces brain reorganizations.
Specifically, interhemispheric connectivity between primary somatosensory areas got closer to a ‘physiological level’ in parallel
with the acquisition of more accurate hand control.

Keywords: Chronic stroke, robotic rehabilitation, resting state EEG, primary somatosensory hand area (S1), interhemispheric
coherence, Functional Source Separation (FSS)

1 These authors contributed equally to this work.

∗ Corresponding author: Giovanni Pellegrino, Department of
Neurology, Campus Bio-Medico University, Via A. del Portillo,
00128 Rome, Italy. Tel.: +39 06 225411286; Fax: +39 06 225411955;
E-mail: g.pellegrino@unicampus.it.

0922-6028/12/$27.50 © 2012 – IOS Press and the authors. All rights reserved
498 G. Pellegrino et al. / Inter-hemispheric coupling changes associate
1. Introduction
Stroke is the third cause of mortality and the first
of disability. Annually, 15 million people worldwide
suffer from a stroke, and it has been evaluated that in
the next years the health burden due to this disease will
rapidly increase. Despite many different interventions
aimed to ameliorate patients’ clinical condition, 30%
or more among all stroke survivors are significantly
disabled due to a neurological impairment involving
force, motion, sensory perception and sensory-motor
integration (Roger et al., 2011).
Among mechanisms sustaining clinical recovery,
cerebral or synaptic plasticity, defined as the neu-
ral ability to change brain functional organization
over time producing different responses to the same
stimulus (Tecchio et al., 2007), seems to play a cru-
cial role (Cramer, 2008; Cramer et al., 2011; Rossini
et al., 2003). The neuroplastic changes linked to a
stroke lesion in the middle cerebral artery (MCA) ter-
ritory leading to a sensory/motor impairment of the
upper limb are able to modulate activity, position and
representation of hand sensorimotor areas bilaterally
(Rossini et al., 2003; Tecchio et al., 2007). Moreover,
such changes affect interhemispheric interactions as
shown by evaluating both cortical excitability through
transcranial magnetic stimulation (TMS) (Grefkes
et al., 2008; Shimizu et al., 2002; Takeuchi et al., 2010;
Traversa et al., 1998) and brain metabolism and acti-
vation through functional magnetic resonance imaging
(fMRI) (Grefkes et al., 2008).
A powerful method that gives us the opportunity to
investigate sensorimotor plasticity is the evaluation at
rest of brain connectivity between ipsilesional hemi-
sphere (ILH) and contralesional hemisphere (CLH)
brain areas reactive to median nerve stimulation
(Wikstrom et al., 2000; Rossini et al., 2004). This tech-
nique is repeatable and not influenced by the patient’s
degree of motor impairment or attention (Allison et
al., 1991); it provides the same input amount to both
hemispheres, and is able to track sensorimotor regions,
thanks to an adhoc procedure named Functional Source
Separation (FSS) (Porcaro et al., 2008).
The FSS technique models the set of EEG signals as
a linear combination of several sources. FSS identifies
a single source at a time, building a contrast function for
that source that exploits some ‘fingerprint’ information
typical of the neuronal pool to be identified. FSS was
used to identify hemispheric cortical sources devoted
to hand representation within S1 (Barbati et al., 2006;
Porcaro et al., 2008; Porcaro et al., 2009), using as fin-
gerprint information the earliest response to the median
nerve stimulation that is known to be generated within
S1 area 3b. Once the sources are identified, FSS allows
us to investigate them in other experimental conditions.
Here, S1 cortical patches in the two hemispheres were
tracked while the subject was at rest.
Furthermore, in patients with a chronic stroke lesion
in the MCA territory and sensorimotor hand impair-
ment it has been demonstrated that primary sensory
cortical areas change their responsiveness to median
nerve stimulation and their topographical organization
involving regions usually not reached by a dense sen-
sory input from the opposite hand (Forss et al., 1999;
Rossini et al., 1998; Rossini et al.; 2001). These plastic
mechanisms are linked to the hand functional recovery
(Altamura et al., 2007; Rossini et al., 1998; Rossini et
al., 2001; Tecchio et al., 2007). FSS allows to track
resting activity of such newly recruited areas devoted
to hand somatosensory perception.
Several studies clearly documented that the neuronal
reorganization occurring after an ischemic insult may
be positively influenced by motor practice, somatosen-
sory input, pharmacological agents (Dimyan and
Cohen, 2011; Laufer and Elboim-Gabyzon, 2011;
Nudo et al., 1996) and neurorehabilitation techniques
including the most innovative ones (Cramer, 2008;
Dimyan and Cohen, 2011). In this sense, specific
robotic systems have been developed as a new and
promising tool to improve motor recovery follow-
ing stroke (Brochard et al., 2010). Even if their
effectiveness compared to standard rehabilitation is
still under debate (Kwakkel et al., 2008), the use of
robotic devices shows several advantages: the abil-
ity to provide a repeatable, programmable, controlled
rehabilitation and the capability of assessing motor fea-
tures in a quantitative manner (Volpe et al., 2009).
Recently, Lo and colleagues (Lo et al., 2010) con-
ducted a multicentric, randomized, controlled trial
on stroke people with long-term disability, to test
the effectiveness of robot- based rehabilitation pro-
gram: the benefit obtained was comparable to that
obtained when a therapist provided intensive therapy.
The study indirectly proved that robot-assisted rehabil-
itation procedures, even in a chronic phase of stroke,
are capable of inducing plasticity sustaining clinical
improvement.
The aim of the present study is to evaluate clinical
changes and brain plastic reorganizations induced by
a robot-aided rehabilitation program using InMotion2
 G. Pellegrino et al. / Inter-hemispheric coupling changes associate 499

Fig. 1. Experimental setup. Left: Robot employed for the rehabilitation program. Centre: Experimental Flow- chart. Right: Hand motor control
and brain activity and connectivity evaluations applied pre- and post-robotic rehabilitation program.

and InMotion3 robots in a population of chronic stroke
patients.
2. Materials and methods
The study was approved by the local Ethical Com-
mittee and informed written consent from all subjects
was obtained.
2.1. Patients
We enrolled seven right-handed patients (age
60 ± 18y, 5 males), with an upper limb sensory/motor
impairment who had suffered a first ever ischemic
stroke in the MCA territory (5 right side, 2 left side).
Time since stroke ranged from 1 to 5 years. In order
to assess the stability of clinical conditions the Upper
Limb Fugl-Meyer Assessment Scale (FMA) (Sivan
et al., 2011) was administered once every six weeks
for three times before the start of the study. MRI
examination confirmed the diagnosis and provided
lesion site characteristics. Patients affected by periph-
eral neuropathy, dementia, severe aphasia or with an
impairment that could bias the correct execution of the
task were excluded.
At the beginning (Tpre ) and at the end (Tpost ) of the
12-week upper-limb robot-aided neurorehabilitation
program all patients underwent a multi-modal eval-
uation that included the assessment of clinical status,
quantitative motor performance and neurophysiologi-
cal features (Fig. 1).
2.2. Clinical and hand motor control evaluation
The patients’ clinical state was evaluated through
the FMA before and after robotic therapy . Moreover,
in order to asses motor performance, we selected a
physiological and simple motor task: the isometric
opposition of the thumb to index and middle fin-
ger against resistance of a semicompliant specifically
designed device. The module used was the ALLADIN
Finger Device (AFD) which consists of a rigid
hand orthesis for isometric tasks, embedding three
force/torque sensors (JR3 model No. 50M31A-I25)
that are located on the outer side of the hand. During the
measurement the hand is positioned between the sen-
sor for the thumb and the two sensors for the index and
middle finger, while the forearm is resting on the arm
support. The fingers are attached to the finger support
using Velcro straps (Figs. 2–3). Visual feedback was
provided by means of a purposely developed graphical
500 G. Pellegrino et al. / Inter-hemispheric coupling changes associate

Fig. 2. ILH and CLH sensorimotor areas variables. Centre: Spatial positioning according to the international 10–20 system of the electrodes
considered for the EEG bipolar derivations overlying the middle cerebral artery territory sensorimotor districts (MCA SM, red) and Primary
Somatosensory Sources (S1, blue) localization for an indicative patient. Left and Right: Power spectral densities (PSD) and Interhemispheric
Coherence (IHCoh) of the Ipsilateral Hemisphere (ILH) and of the Contralateral Hemisphere (CLH) obtained from both Primary Somatosensory
Sources (S1) and EEG bipolar derivations overlying the middle cerebral artery territory sensorimotor districts (MCA SM) at the beginning
(Tpre ) and at the end (Tpost ) of the rehabilitation program for an indicative patient. For MCA CM, PSDs were separately calculated in the
ipsilesional (ILH) and contralesional hemisphere (CLH) averaging the bipolar derivations of each hemisphere. The inter-hemispheric coherence
was calculated with a similar procedure. (Colours are visible in the online version of the article; http://dx.doi.org/10.3233/RNN-2012-120227)

interface to control that the applied pressure matched
the required level of force. Initially the subjects were
asked to perform the task at maximal voluntary con-
traction (MVC) force with each hand separately, to
calculate their residual ability. After a rest period of
at least 5 minutes, in order to avoid fatigue in the test
task, subjects alternated 20 seconds of isometric con-
traction to 20 seconds of rest for 15 times repeatedly
with each hand. The target force level was set to 20%
of MVC and for each hand about 300 s of contrac-
tion were recorded. Paretic and non-paretic hand motor
control abilities were estimated by: 1) MVC; 2) the
level of applied force along the whole task (average of
the applied force during all contraction periods, Con-
traction level); 3) an index incorporating the capability
to apply a strength within ±5% of the established level
of contraction and its duration (the product of time and
mean force level, Contraction quality).
2.3. Robotic rehabilitation
All patients underwent a 12-week robotic rehabil-
itation program with the shoulder- and -elbow and
wrist robotic machines, namely the InMotion2 (or
MIT-Manus) and the InMotion3 robots, respectively
(Zollo et al., 2011; Zollo et al., 2011). The MIT-
Manus machine is a planar, two degree-of-freedom
robot providing assistance to the patient upper extrem-
ity motion while executing a series of “video games”
that involve positioning the robot end effector (Krebs
et al., 2007). The wrist robot, instead, has three
active DOF: abduction–adduction; flexion–extension;
pronation–supination. The two side-mounted actua-
tors are connected to a differential mechanism, which
enables flexion/extension, abduction/adduction move-
ments and their combinations. Joint torque production
on the differential mechanism is a result of the
proper combination of motor torques. When the two
motors rotate in the same direction, motion is purely
abduction/adduction; when they rotate in the oppo-
site direction, the resulting motion is flexion/extension.
The entire differential mechanism is mounted onto a
curved rack so that it can be actuated from beneath
the forearm, enabling pronation and supination move-
ments (Krebs et al., 2007). Thus, a third motor is used
to actuate the pronation/supination degree of freedom.
Each subject has been trained 1 hour a day for three
days a week. All subjects underwent 6 weeks of train-
ing with the InMotion2 robot and 6 weeks with the
InMotion3 robot.
 G. Pellegrino et al. / Inter-hemispheric coupling changes associate 501

MAXIMAL VOLUNTARY CONTRACTION (MVC) CONTRACTION LEVEL CONTRACTION QUAUTY

60 * 15 * ** **
* **
** * 6

** **

PRE POST PRE POST PRE POST

Robotic rehabilitation Robotic rehabilitation Robotic rehabilitation

Moved hand

Paretic Non Paretic

Fig. 3. Robotic rehabilitation effects on hand motor control. Top: Error bars indicate mean ±1 SE for the three indices considered for hand
motor control assessment: Maximal Voluntary Contraction (MVC), Contraction Level and Contraction Quality. X-axis PRE and POST Robotic
Rehabilitation. Y-axis: For Maximal Voluntary Contraction (MVC) and Contraction Level, strength applied expressed in N; for execution quality
strength within ±5% of the established level of contraction and its duration (N*sec). Blue Bars: Paretic Hand; Red Bars: Non Paretic Hand.
Significant differences: *corresponds to p < 0.05; **corresponds to p < 0.001. Bottom: Alladin device (finger module) employed for Hand Motor
Control measures of the Paretic and Non Paretic Hand. (Colours are visible in the online version of the article; http://dx.doi.org/10.3233/RNN-
2012-120227)

2.4. Neurophysiological investigation and during median nerve stimulation. Stimulation

2.4.1. EEG recordings
Thirty-two channel EEG (Fp1, Fp2, F3, F4, C3,
C4, P3, P4, O1, O2, F7, F8, P7, P8, T7, T8, FZ,
CZ, PZ, FC1, FC2, CP1, CP2, FC5, FC6, FT9, FT10,
FCZ, CP5, CP6, TP9, TP10) with binaural refer-
ence was recorded with scalp electrodes mounted on
an elastic cap, according to the 10–20 international
system. Vertical and horizontal electro-oculogram
were recorded bipolarly in order to control for eye
movement-related artifacts. Impedances of all elec-
trodes were kept below 5 k
. EEG data were sampled
at 1024 Hz (pre-sampling analogical band-pass filter
set at 0.48–256 Hz, BrainAmp System). EEG record-
ing was performed at rest with eyes opened (5 minutes)
consisted in short electric pulses (0.2 ms duration,
intensity set at three times the subjective sensory
threshold producing a painless clearly visible thumb
opposition, and interstimulus interval of 631 ms) deliv-
ered unilaterally to the right and left median nerve at
the wrist. Stimuli were delivered through a pair of non-
magnetic, 2.5-cm spaced, Ag–AgCl disc electrodes
filled with conductive jelly.
2.4.2. Primary sensorimotor hand area
identification
Two different methods were applied to identify the
sensorimotor area of the ILH and CLH: a) EEG bipolar
derivations overlying the (MCA) territory sensorimo-
tor districts (MCA SM) considering scalp EEG (for
502 G. Pellegrino et al. / Inter-hemispheric coupling changes associate
the left hemisphere: F3-C3; C3-P3; FC1-CP1; FC5-
CP5 and the homologous sites for the right side).
b) the somatosensory source extracted from the 32
EEG signals by applying the previously mentioned
FSS procedure (Porcaro et al., 2009; Tecchio et al.,
2007).
FSS models a set of signals X a linear combination
(through an unknown mixing matrix A) of sources S:
X = AS
The single source FS is obtained optimizing the con-
trast function:
F = J + λR
where J can be any function normally used for Inde-
pendent Component Analysis (in our case kurtosis),
while H accounts for the prior information we have on
sources. Parameter λ is used to suitably weigh the two
parts of the contrast function.
To identify neural network devoted to somatosen-
sory perception from the hand (territory supplied by
the median nerve), the ‘reactivity’ to the stimuli was
taken into account. It was defined as follows: the
evoked activity (EA) was computed by averaging sig-
nal epochs centered on the stimulus of the median
nerve.
The reactivity coefficient (R) was then computed as:

40
10
− hemispheric coherence was calculated with a similar
R = |EA(t)| dt − |EA(t)| dt
20 −30
with t = 0 corresponding to the stimulus delivery to the
median nerve at wrist. The time interval ranging from
20 to 40 ms includes the maximum activation (Tec-
chio et al., 1997) and the baseline (no response) was
computed in the pre-stimulus time interval (−30 to
−10 ms).
The FSS algorithm was applied to EEG recordings
during contralateral median nerve stimulation to iden-
tify time signal and field distribution of bilateral S1
neuronal pools. EEG channel weights, in turn, allowed
position identification of the two sources by applying
proper inverse problem algorithms. To investigate right
and left S1 at rest, a semi-automatic artifact rejection
procedure (Barbati et al., 2004) was applied to EEG
data recorded in a resting state to minimize the con-
tribution of non-cerebral sources (such as the heart,
eyes, muscles) which can critically exceed the brain
signal in absence of stimulus synchronized average
noise-reduction. Thereafter, artifact-free rest EEG data
were multiplied by the inverse of the demixing matrix
of S1 (WS1 = 1/AS1) to obtain activity in the resting
state.
Once the source extraction was completed, the FSS
method allowed to evaluate features such as activity,
connectivity and position of the two primary cortical
neuronal pools devoted to the contra-lateral hand (S1,
ILH S1 for ipsilesional hemisphere and CLH S1 for
(1) contralesional hemisphere), also at rest.
The S1 source localization was suitably obtained by
applying the sLoreta algorithm (http://www.uzh.ch/
keyinst/NewLORETA/Methods/MethodsSloreta.htm)
(2) (Pascual-Marqui, 2002). The source position
coordinates were expressed in the Talairach system.
2.4.3. Sensorimotor area properties at rest
Hemispheric power spectral density (PSD) and
inter-hemispheric coherence (IHCoh) were estimated
for both S1 and MCA SM at the resting state (Fig. 2).
The PSD was calculated through a standard FFT
approach using Welch technique and a Hanning win-
dowing function (window 2048 ms, no overlap). The
IHCoh was calculated using a fixed number of 120
averaged trails.
S1 PSDs were calculated separately in ILH and
CLH, and MCA SM PSDs were obtained averaging
each hemisphere’s bipolar derivations. The inter-
(3) procedure. PSD frequency bands were defined on
the basis of the individual alpha frequency (IAF):
delta (2, 3.5), theta (4, IAF−2.5 Hz), alpha (IAF−2,
IAF+2 Hz), low beta (IAF+2, 23.5 Hz), high beta
(24–32.5 Hz), low gamma (33–47.5 Hz), high gamma
(52.5, 90 Hz). Each band power was normalized to the
total power in the (2, 90 Hz) frequency range.
2.5. MRI
Brain MRI was carried out at 1.5 T, using spin-
echo (SE), turbo spin-echo (TSE), and fluid-attenuated
inversion recovery (FLAIR) sequences. All sequences
provided contiguous 5-mm-thick slices on sagittal,
coronal, and axial planes. Lesion characterization was
performed on axial slices. Lesions were classified as
“cortical” (C) when mainly the cortical gray mat-
ter was involved or “subcortical” (S) when there
was no visible cortical involvement and white mat-
ter, internal capsule, basal ganglia, and thalamus were
affected.
 G. Pellegrino et al. / Inter-hemispheric coupling changes associate
2.6. Statistical analysis
After checking the frequency distribution of each
measure, suitable transformations, in order to achieve
a better approximation to gaussianity and a good
control of outliers, were applied if necessary. The
main statistical analysis aimed at assessing the
effects of robot-aided rehabilitation on clinical patient
conditions, in terms of both FMAS scores and afore-
mentioned motor performance indices, and on the
cerebral reorganization possibly paralleling the clini-
cal changes. In agreement with single FMAS scores
in each condition for each patient, paired samples
t-tests were performed. Since motor performance
indices for each patient were obtained in several tri-
als (15) for the movement of both the paretic and
non-paretic hand and in pre- and post- robotic reha-
bilitation evaluations, we decided not to collapse all
the available information, thus avoiding consider-
ing the averages of the 15 values. We therefore did
not apply a traditional ANOVA for repeated mea-
sures. Instead, we considered all data points and
applied a General Estimating Equation model, with
the patient as the “Subject” (or cluster) variable,
motor performance values as the dependent variable
and Moved hand (paretic, non-paretic) and Robotic
rehab (pre-, post-robotic rehabilitation) as predic-
tors. Since motor performance indices were estimated
on consecutive epochs, we chose the autoregressive
(lag = 1) as the working correlation within subjects.
The alpha-inflation due to multiple comparisons was
faced according to Sidak’s procedure (similar to Bon-
ferroni’s procedure but slightly more powerful). When
evaluating cerebral reorganizations in terms of S1 posi-
tions and responsiveness to median nerve stimulation,
spectral features and inter-hemispheric coherence of
primary somatosensory areas and cortical regions sup-
plied by the middle cerebral artery similar models
were applied. Predictors were Robotic rehab (pre-,
post- rehabilitation) and Hemisphere (ILH, CLH). Fre-
quency band (delta, theta, alpha, low beta, high beta,
low gamma, high gamma) was added for spectral pow-
ers and coherences.
2.6.1. Relationship between cerebral variables
and performance and clinical indexes
The neurophysiological measures were studied
to investigate their functional relevance in relation
to motor performance before and after robotic reha-
bilitation, and to clinical recovery levels. Bivariate
503
parametric correlation analysis was performed
between neurophysiological measures impacted by
robot-aided rehabilitation and the clinical conditions
of patients (FMA and motor performance).
3. Results
3.1. Effects of robot-aided rehabilitation: Clinical
aspects
NIHSS at stroke onset was 8.4 ± 3.0. Patients suf-
fered a moderate to severe residual impairment before
being admitted to the robotic rehabilitation program.
Patients improved their clinical state as indicated by
ameliorations of both shoulder (15.0 ± 7.2 at Tpre
17.6 ± 7.0 at Tpost (p = 0.007) and wrist items (Tpre
16.8 ± 12.4 and Tpost 19.4 ± 13.8, p = 0.041). MRI
examination provided lesion site characteristics, show-
ing that six patients had a subcortical lesion while one
only had a cortical lesion.
3.2. Effects of robot-aided rehabilitation: Motor
hand control
The ability to perform the experimentally required
motor task was improved by the robotic rehabilitation
intervention.
3.2.1. Maximal voluntary contraction (MVC)
A strong interaction Moved hand*Robotic rehab
effect [Wald chi-square = 8.159, df = 1, p = 0.004]
revealed that the changes were more evident for the
rehabilitated paretic hand [Fig. 3, post-hoc analysis
Sidak’s corrected p = 0.001] than for the non-paretic
hand [Fig. 3, p = 0.050]. Furthermore, we observed that
the significant disparity in the ability to execute the task
between the two hands at Tpre [Fig. 3, p < 0.001] was
reduced at Tpost [p = 0.035].
3.2.2. Contraction level
Patients were able to maintain a higher level of con-
traction at Tpost compared to Tpre [Fig. 3, Robotic rehab
factor Wald chi-square = 6.775, df = 1, p = 0.009].
Interestingly, the improvement was observed bilat-
erally, as indicated by the absence of interaction
Moved hand*Robotic rehab [p = 0.974]. Nonetheless
and needless to say, at both Tpre and Tpost the non-
paretic hand performed better than the paretic one
504 G. Pellegrino et al. / Inter-hemispheric coupling changes associate
[Moved hand effect Wald chi-square = 52.893, df = 1,
p < 0.001].
3.2.3. Contraction quality
A strong interaction Moved hand*Robotic rehab
effect [Wald chi-square = 6.824, df = 1, p = 0.009]
revealed that the execution quality of the test task
highly increased for the rehabilitated paretic hand
[Fig. 3, post-hoc analysis Sidak’s corrected p < 0.001]
and not at all for the non-paretic hand [Fig. 3,
p = 0.711]. Noteworthy, paretic hand execution quality
become similar to the non-paretic hand after rehabili-
tation [post-hoc analysis Sidak’s corrected p = 0.571],
while it was much worst before [p < 0.001].
3.3. Effects of robot-aided- rehabilitation:
Sensorimotor hand representation
3.3.1. S1 position and responsiveness to median
nerve stimulation
Robotic rehabilitation did not impact on the posi-
tion of S1 source in ILH or CLH [Robotic rehab and
interaction factors p > 0.200 consistently].
3.3.2. S1 resting state activity
No significant changes in resting state activities
linked to robotic rehabilitation were found.
3.3.3. Resting state functional connectivity
between ILH and CLH S1s
Robotic rehabilitation impacted the level of
inter-hemispheric coherence between S1 [Robotic
rehab factor Wald chi-square = 4.971, df = 1,
p = 0.026, Figs. 4–5]. This change after the reha-
bilitation period depended on frequency content
[Robotic rehab*Frequency band factor Wald chi-
square = 54.575, df = 6, p < 0.001, Figs. 4–5]. IHCoh
reduction after rehabilitation was evident in high beta
and gamma bands [high beta p = 0.027, low gamma
p = 0.019, high gamma p = 0.015; Figs. 4–5]. Note-
worthy, the inter-hemispheric S1 coherences were
more spread before than after robotic rehabilitation.
In fact, Levine test on the differences between IHCoh
standard deviations pre- and post-rehabilitation had
significances p = 0.002, 0.015, 0.009, 0.011 for low
and high beta, low and high gamma bands respectively
(Fig. 5). In delta, theta and alpha bands, while the
mean level did not differ between pre- and post-
rehabilitation, the S1 IHCoh values were more spread
pre- than post-rehabilitation (p = 0.008; 0.022; <0.001
respectively. p = 0.019 on the whole spectrum, Fig. 5).
3.3.4. MCA SM territory resting state activity
No significant changes in resting state activities
linked to robotic rehabilitation were found.
3.3.5. Resting state functional connectivity
between ILH and CLH MCA SM
Robotic rehabilitation did not modify the lev-
els of coherence among the neuronal activities of
ILH and CLH regions by MCA. We noted that
the inter-hemispheric coherence was hugely higher
between ILH and CLH S1s than between whole
homologous regions supplied by MCAs [S1s about
46 times MCA territory regions, Fig. 4; Cerebral
Region (S1, MCA SM) Wald chi-square = 16.273,
df = 1, p < 0.001].
3.4. Relationship between robotic rehabilitation
effects on inter-hemispheric sensorimotor
connectivity and hand motor control
Hand control changes after robotic rehabilitation
were correlated with changes of brain sensorimotor
organization.
3.4.1. Inter-hemispheric coherence
Contraction level of the paretic hand improved after
robotic rehabilitation in strict correlation with pre-
and post- rehabilitation S1 IHCoh changes [r = 0.758,
p = 0.048 in high beta band, Fig. 6]. The non-paretic
hand Contraction level displayed a similar trend
[r = 0.739, p = 0.058 in high beta band, Fig. 6]. No
significant effects were observed for MCA SM IHCoh.
4. Discussion
Upper-limb robotic rehabilitation produces inter-
hemispheric connectivity changes associated with
improved task performance.
The main evidence of the present study is that 12
weeks of upper limb robotic rehabilitation treatment
in chronic stroke patients change the inter-hemispheric
functional connectivity between primary sensorimotor
hand areas in correlation with the paretic hand motor
control improvement.
We found that the inter-hemispheric connectivity
was distributed across a wide range before robotic
 G. Pellegrino et al. / Inter-hemispheric coupling changes associate 505

Fig. 4. Robotic rehabilitation effects on inter-hemispheric sensorimotor connectivity. Left: Interhemispheric coherence (IHCoh) for S1 and
MCA SM regions at Tpre and Tpost . Error bars ± 1 SE. S1 IHCoh levels are 46 times higher than MCA SM IHCoh. Frequency bands defined on
the basis of the individual alpha frequency (IAF): delta (2, 3.5), theta (4, IAF−2.5 Hz), alpha (IAF−2, IAF+2 Hz), low beta (IAF+2, 23.5 Hz),
high beta (24–32.5 Hz), low gamma (33–47.5 Hz), high gamma (52.5, 90 Hz). Significant differences: *refers to S1 IHCoh difference between
Tpre 4 and Tpost and corresponds to p < 0.05; Right: Larger view of MCA SM IHCoh. There are no significant differences between Tpre and Tpost .

Fig. 5. S1 IHCoh values distribution before and after robotic rehabilitation. S1 IHCoh values distribution (colored dots) and level (empty dots,
mean ± 1SE) at the beginning and at the end of the rehabilitation program. At the top of the figure, the upper line refers to Levene test applied
to Tpre and Tpost values distribution, the lower one refers to Tpre and Tpost S1 IHCoh difference. Note that the inter-hemispheric coherence
was distributed across a wide range before robotic rehabilitation and distributed tightly to a final level after treatment. Significant differences:
*corresponds to p < 0.05; **corresponds to p < 0.001.

rehabilitation and distributed tightly to a final level found at the end of the rehabilitation program corre-
after treatment. Since a clear motor improvement sponds to a more functionally efficacious and ‘physio-
induced by robotic rehabilitation was observed, it could logical’ condition. Previous findings on improvements
be speculated that the inter-hemispheric connectivity after robot-aided rehabilitation in stroke patients
506 G. Pellegrino et al. / Inter-hemispheric coupling changes associate

Fig. 6. Relationship between robotic rehabilitation effects on inter-hemispheric sensorimotor connectivity and hand motor control. Left: S1
IHCoh distributions and levels at the beginning and at the end of the rehabilitation program (High beta band has been chosen as indicative).
Centre: Contraction Level distributions and levels at the beginning and at the end of the rehabilitation program. Both Paretic and Non Paretic
Hand improve their hand motor control. Right: Graphical representation of correlations between the change of S1 interhemispheric coherence
(Tpost -Tpre , X Axis) and the improvement of Contraction Level (Tpost -Tpre , Y Axis).

(Brochard et al., 2010) are confirmed in the present
study; they are also expanded by the demonstration that
a robot-based rehabilitation program is able to induce
brain reorganizations and that inter-hemispheric
connectivity between primary somatosensory areas
changed as more fine hand control was achieved.
4.1. FSS enabled the description of a specific
network at rest
Previous EEG studies on cortical connectivity in
stroke patients used task–related protocols and showed
reorganizations with a drastic reduction of information
flow and the engagement of only few cerebral areas in
beta and gamma bands (de Vico Fallani et al., 2009).
In well recovered chronic stroke patients, task-
related EEG connectivity exhibits a shift mainly
towards contralesional sensorimotor areas, showing
an increase of low beta inter- hemispheric coher-
ence between sensory-motor regions (Gerloff et al.,
2006). The increased coupling between cortical areas
in particular in beta band (Wheaton et al., 2008) may
represent a compensatory mechanism secondary to
stroke and an interhemispheric coupling decrease can
occur as patients make a functional recovery (Strens
et al., 2004). At difference from previous studies, in
the present investigation brain connectivity between
the cortical areas devoted to the paretic and non-
paretic hand perception was investigated at resting
state. Recent evidences suggest that the networks
related to specific functional domains show relevant
functional properties evident in the resting state (Deco
and Corbetta, 2011). It has been suggested that evalu-
ating functional connectivity in resting state could be
useful for assessing the health of brain networks, with
crucial functional importance of inter-hemispheric
interactions (Carter et al., 2010). Moreover, a defi-
nite advantage of functional connectivity study during
resting state, especially in the sensorimotor function
assessment of stroke patients, is to get rid of dif-
ferent task-related neural engagements (i.e., attention
and monitoring in planning and executing the motor
task with a paretic hand), not comparable between
patients and controls and between paretic and non-
paretic hands. In this respect the use of FSS to evaluate
functional connectivity in resting state for the spe-
cific regions devoted to the hand sensorimotor control
independently of their location was extremely helpful.
 G. Pellegrino et al. / Inter-hemispheric coupling changes associate
Since functional sources (S1) are identified on the base
of their specific response to the hand stimulation, they
can be suitably investigated although sensorimotor
areas displacements may occur due to plastic reorgani-
zation phenomena following ischemic lesion (Tecchio
et al., 2006, Tecchio et al., 2007). Interestingly, we
found that robotic treatment improved hand control
ability in patients and modulated the inter-hemispheric
connectivity in resting state. In this sense we hypoth-
esized that a causative correlation between changes
of interhemispheric coupling and clinical outcome
may be possible. The inter-hemispheric coherence
parameter is not able to define whether the functional
engagement of the bilateral sensory areas is linked
to excitatory or inhibitory activity. It has been well
demonstrated that local inhibitory networks receive
major transcallosal projections, mediating the inter-
hemispheric connectivity (Lee et al., 2007); thus, it
could be argued that hemispheric inhibitory structures
are involved in the functional connectivity between
homologous cortices in the two hemispheres. Further-
more, an imbalance of sensorimotor inter-hemispheric
interaction, in which the affected hemisphere is exces-
sively inhibited by the unaffected hemisphere, has been
extensively demonstrated (Murase et al., 2004; Oliveri
et al., 1999; Traversa et al., 1998). Both motor (Shimizu
et al., 2002; Takeuchi et al., 2010) and somatosen-
sory cortex (Floel et al., 2004) are involved in this
phenomenon. Neuromodulation interventions are set
up exploiting this concept: inhibitory rTMS applied
over the unaffected hemisphere, aiming at reducing its
depressive influence on the affected one, produces an
improvement of motor function (Mansur et al., 2005)
and reduces hyper-activations of the unaffected hemi-
sphere (Nowak et al., 2008), increases the activation
of the affected hemisphere (Conchou et al., 2009) and
induces a reduction of interhemispheric effective con-
nectivity paired to clinical improvement (Grefkes et al.,
2010; Landi and Rossini, 2010; Rossini et al., 2003).
Conceivably in our case, rearrangements of excitatory-
inhibitory local and interhemispheric balances subtend
the modulation of functional connectivity between
affected and unaffected sensory hand areas.
4.2. Inter-hemispheric connectivity is crucial for
behaviour
Brain plastic reorganizations occurring after an
ischemic stroke can contribute to functional recovery
(Cramer et al., 2011; Rossini et al., 2003; Tecchio
507
et al., 2006; Tecchio et al., 2007). In parallel with
regional activation changes, it is well known that the
modulation of brain areas connectivity plays a crucial
role (Grefkes and Fink, 2011; Siegel et al., 2012). The
level of brain connectivity depends on the epoch of
evaluation from stroke onset (Wang et al., 2010), on
the condition studied (rest vs. task related) and on
the technique applied for the evaluation (Muthuku-
maraswamy and Singh, 2008). In animal models (van
Meer et al., 2010) and humans (Carter et al., 2010) low
inter-hemispheric connectivity are typically observed
in acute phase coupled to severe clinical state, while
progressive recovery of motor control is paired to an
increase of sensorimotor homologous areas connectiv-
ity. In this framework in acute stroke patients studied
longitudinally along one year, Wang et al. (Wang et al.,
2010) showed that the resting state inter-hemispheric
coupling between fMRI BOLD signals of homologous
primary motor cortices played a crucial role for motor
improvements. Overall, the resting state fMRI investi-
gations suggest that the interaction between different
brain areas enter the recovery of lost functions due to
a stroke lesion, with both increases and decreases of
functional coupling marking both adaptive and mal-
adaptive phenomena (Grefkes and Fink, 2011; Wang
et al. 2010).
4.3. Sensitivity of source identification procedure
We measured the connectivity between affected and
un-affected sensorimotor regions, either by consider-
ing the EEG activity in the central region supplied
by the middle cerebral artery (MCA SM), or focusing
on the somatosensory sources devoted to the paretic
and the non-paretic hand (S1, by FSS algorithm (Por-
caro et al., 2009)). Exclusively the coherence between
S1 hand sources activities was impacted by robot-aided
rehabilitation, selectively in high beta and gamma
bands. The level of connectivity between S1s resulted
46 times higher than between MCA SMs, conceivably
since FSS method is able to more accurately iden-
tify the primary sensory neuronal pools devoted to
hand representation. In fact, this investigation strat-
egy allowed to clearly reveal the inter- hemispheric
coupling modulation between homologous areas that
was not evident considering the whole MCA SM ter-
ritory signals. Further advantage of FSS method is to
better catch high frequency rhythms than EEG scalp
activity (Porcaro et al., 2009). Thanks to this proce-
dure, significant changes in beta and gamma bands
508 G. Pellegrino et al. / Inter-hemispheric coupling changes associate
coherence subsequent to robotic rehabilitation have
been recognized. The involvement of beta band syn-
chronization in sensory motor integration has been
well demonstrated (Engel and Fries, 2010; Tombini
et al., 2009). Nonetheless, gamma-band connectivity
plays a role for the dynamic construction of small- and
large-scale neuronal networks related to sensorimotor
integration (Alonso et al., 1996; Szurhaj and Deram-
bure, 2006; Tecchio et al., 2007; Tecchio et al., 2008).
Consistent with these data we found that the modifica-
tion of inter-hemispheric connectivity in resting state
associated with hand control improvements occurred
in frequency ranges crucial for active motor control.
5. Conclusion
The FSS method, fully independent on patient’s
compliance, allows to highlight plastic changes linked
with motor improvements and adds new evidence
that resting state inter-hemispheric connectivity could
impact behavioral control. Finally, our study demon-
strates that robotic rehabilitation obtains in chronic
stroke patients a modulation of resting state func-
tional connectivity between homologous sensorimotor
regions directed toward an ‘in-between’ more effective
level linked to better motor performance.
Acknowledgments
Authors would like to thank Professor PM Rossini
for the inspiration of this study, his encouragement and
his continuous support and NFPT Rita Fini for excel-
lent technical assistance. The research leading to these
results has received funding from 1. Italian Ministry of
Health Cod. GR-2008-1138642 ‘Promoting recovery
from Stroke: Individually enriched therapeutic inter-
vention in Acute phase’ [ProSIA]; 2. Royal Society
International Joint Project - 2010/R1: ‘The Key Move-
ment Controllers: an EEG/fMRI study of the hand
network dynamics [KeyMoCo]’.
References
Allison, T., McCarthy, G., Wood, C.C. & Jones, S.J. (1991).
Potentials evoked in human and monkey cerebral cortex by
stimulation of the median nerve. A review of scalp and intracra-
nial recordings. Brain, 114, 2465-2503.
Alonso, J.M., Usrey, W.M. & Reid, R.C. (1996). Precisely correlated
firing in cells of the lateral geniculate nucleus. Nature, 383,
815-819.
Altamura, C., Torquati, K., Zappasodi, F., Ferretti, A., Pizzella, V.,
Tibuzzi, F., Vernieri, F., Pasqualetti, P., Landi, D., Del Gratta,
C., Romani, G.-L., Maria Rossini, P. & Tecchio, F. (2007).
fMRI-vs-MEG evaluation of post-stroke interhemispheric
asymmetries in primary sensorimotor hand areas. Exp Neurol,
204, 631-639.
Barbati, G., Porcaro, C., Zappasodi, F., Rossini, P.M. & Tecchio,
F. (2004). Optimization of an independent component analysis
approach for artifact identification and removal in magnetoen-
cephalographic signals. Clin Neurophysiol, 115, 1220-1232.
Barbati, G., Sigismondi, R., Zappasodi, F., Porcaro, C., Graziadio,
S., Valente, G., Balsi, M., Rossini, P.M. & Tecchio, F. (2006).
Functional source separation from magnetoencephalographic
signals. Hum Brain Mapp, 27, 925-34.
Brochard, S., Robertson, J., Medee, B. & Remy-Neris, O. (2010).
What’s new in new technologies for upper extremity rehabili-
tation? Curr Opin Neurol, 23, 683-687.
Carter, A.R., Astafiev, S.V., Lang, C.E., Connor, L.T., Rengachary,
J., Strube, M.J., Pope, D.L.W., Shulman, G.L. & Corbetta,
M. (2010). Resting interhemispheric functional magnetic reso-
nance imaging connectivity predicts performance after stroke.
Ann Neurol, 67, 365-375.
Conchou, F., Loubinoux, I., Castel-Lacanal, E., Le Tinnier, A.,
Gerdelat-Mas, A., Faure-Marie, N., Gros, H., Thalamas, C.,
Calvas, F., Berry, I., Chollet, F. & Simonetta Moreau, M. (2009).
Neural substrates of low-frequency repetitive transcranial mag-
netic stimulation during movement in healthy subjects and acute
stroke patients. A PET study. Hum Brain Mapp, 30, 2542-2557.
Cramer, S.C. (2008). Repairing the human brain after stroke. II.
Restorative therapies. Ann Neurol, 63, 549-560.
Cramer, S.C., Sur, M., Dobkin, B.H., O’Brien, C., Sanger, T.D., Tro-
janowski, J.Q., Rumsey, J.M., Hicks, R., Cameron, J., Chen,
D., Chen, W.G., Cohen, L.G., deCharms, C., Duffy, C.J., Eden,
G.F., Fetz, E.E., Filart, R., Freund, M., Grant, S.J., Haber, S.,
Kalivas, P.W., Kolb, B., Kramer, A.F., Lynch, M., Mayberg,
H.S., McQuillen, P.S., Nitkin, R., Pascual-Leone, A., Reuter-
Lorenz, P., Schiff, N., Sharma, A., Shekim, L., Stryker, M.,
Sullivan, E.V. & Vinogradov, S. (2011). Harnessing neuroplas-
ticity for clinical applications. Brain, 134, 1591-1609.
de Vico Fallani, F., Astolfi, L., Cincotti, F., Mattia, D., la Rocca, D.,
Maksuti, E., Salinari, S., Babiloni, F., Vegso, B., Kozmann, G. &
Nagy, Z. (2009). Evaluation of the brain network organization
from EEG signals: A preliminary evidence in stroke patient.
Anat Rec (Hoboken), 292, 2023-2031.
Deco, G. & Corbetta, M. (2011). The dynamical balance of the brain
at rest. Neuroscientist, 17, 107-123.
Dimyan, M.A. & Cohen, L.G. (2011). Neuroplasticity in the context
of motor rehabilitation after stroke. Nat Rev Neurol, 7, 76-85.
Engel, A.K. & Fries, P. (2010). Beta-band oscillations–signalling the
status quo? Curr Opin Neurobiol, 20, 156-165.
Floel, A., Nagorsen, U., Werhahn, K.J., Ravindran, S., Birbaumer,
N., Knecht, S. & Cohen, L.G. (2004). Influence of somatosen-
sory input on motor function in patients with chronic stroke.
Ann Neurol, 56, 206-212.
 G. Pellegrino et al. / Inter-hemispheric coupling changes associate
Forss, N., Hietanen, M., Salonen, O. & Hari, R. (1999). Modified
activation of somatosensory cortical network in patients with
right-hemisphere stroke. Brain, 122, 1889-1899.
Gerloff, C., Bushara, K., Sailer, A., Wassermann, E.M., Chen, R.,
Matsuoka, T., Waldvogel, D., Wittenberg, G.F., Ishii, K., Cohen,
L.G. & Hallett, M. (2006). Multimodal imaging of brain reorga-
nization in motor areas of the contralesional hemisphere of well
recovered patients after capsular stroke. Brain, 129, 791-808.
Grefkes, C., Nowak, D.A., Eickhoff, S.B., Dafotakis, M., Kust, J.,
Karbe, H. & Fink, G.R. (2008). Cortical connectivity after sub-
cortical stroke assessed with functional magnetic resonance
imaging. Ann Neurol, 63, 236-246.
Grefkes, C., Nowak, D.A., Wang, L.E., Dafotakis, M., Eickhoff,
S.B. & Fink, G.R. (2010). Modulating cortical connectivity
in stroke patients by rTMS assessed with fMRI and dynamic
causal modeling. Neuroimage, 50, 233-242.
Grefkes, C. & Fink, G.R. (2011). Reorganization of cerebral net-
works after stroke: New insights from neuroimaging with
connectivity approaches. Brain, 134, 1264-1276.
Krebs, H.I., Volpe, B.T., Williams, D., Celestino, J., Charles, S.K.,
Lynch, D. & Hogan, N. (2007). Robot-aided neurorehabilita-
tion: A robot for wrist rehabilitation. IEEE Trans Neural Syst
Rehabil Eng, 15, 327-335.
Kwakkel, G., Kollen, B.J. & Krebs, H.I. (2008). Effects of robot-
assisted therapy on upper limb recovery after stroke: A
systematic review. Neurorehabil Neural Repair, 22, 111-121.
Landi, D. & Rossini, P.M. (2010). Cerebral restorative plasticity
from normal ageing to brain diseases: A “never ending story”.
Restor Neurol Neurosci, 28, 349-366.
Laufer, Y. & Elboim-Gabyzon, M. (2011). Does sensory transcuta-
neous electrical stimulation enhance motor recovery following
a stroke? A systematic review. Neurorehabil Neural Repair, 25,
799-809.
Lee, H., Gunraj, C. & Chen, R. (2007). The effects of inhibitory and
facilitatory intracortical circuits on interhemispheric inhibition
in the human motor cortex. J Physiol, 580, 1021-1032.
Lo, A.C., Guarino, P.D., Richards, L.G., Haselkorn, J.K., Witten-
berg, G.F., Federman, D.G., Ringer, R.J., Wagner, T.H., Krebs,
H.I., Volpe, B.T., Bever, C.T., Bravata, D.M., Duncan, P.W.,
Corn, B.H., Maffucci, A.D., Nadeau, S.E., Conroy, S.S., Pow-
ell, J.M., Huang, G.D. & Peduzzi, P. (2010). Robot-assisted
therapy for long-term upper-limb impairment after stroke. N
Engl J Med, 362, 1772-1783.
Mansur, C.G., Fregni, F., Boggio, P.S., Riberto, M., Gallucci-Neto,
J., Santos, C.M., Wagner, T., Rigonatti, S.P., Marcolin, M.A.
& Pascual-Leone, A. (2005). A sham stimulation-controlled
trial of rTMS of the unaffected hemisphere in stroke patients.
Neurology, 64, 1802-1804.
Murase, N., Duque, J., Mazzocchio, R. & Cohen, L.G. (2004).
Influence of interhemispheric interactions on motor function
in chronic stroke. Ann Neurol, 55, 400-409.
Muthukumaraswamy, S.D. & Singh, K.D. (2008). Spatiotemporal
frequency tuning of BOLD and gamma band MEG responses
compared in primary visual cortex. Neuroimage, 40, 1552-
1560.
Nowak, D.A., Grefkes, C., Dafotakis, M., Eickhoff, S., Kust, J.,
Karbe, H. & Fink, G.R. (2008). Effects of low-frequency
509
repetitive transcranial magnetic stimulation of the con-
tralesional primary motor cortex on movement kinematics
and neural activity in subcortical stroke. Arch Neurol, 65,
741-747.
Nudo, R.J., Wise, B.M., SiFuentes, F. & Milliken, G.W. (1996).
Neural substrates for the effects of rehabilitative training on
motor recovery after ischemic infarct. Science, 272, 1791-
1794.
Oliveri, M., Rossini, P.M., Traversa, R., Cicinelli, P., Filippi, M.M.,
Pasqualetti, P., Tomaiuolo, F. & Caltagirone, C. (1999). Left
frontal transcranial magnetic stimulation reduces contralesional
extinction in patients with unilateral right brain damage. Brain,
122, 1731-1739.
Pascual-Marqui, R.D. (2002). Methods Find Exp Clin Pharmacol.
Methods and Findings in Experimental and Clinical Pharma-
cology, 24(Suppl D), 5-12.
Porcaro, C., Barbati, G., Zappasodi, F., Rossini, P.M. & Tec-
chio, F. (2008). Hand sensory-motor cortical network assessed
by functional source separation. Hum Brain Mapp, 29,
70-81.
Porcaro, C., Coppola, G., Di Lorenzo, G., Zappasodi, F., Siracusano,
A., Pierelli, F., Rossini, P.M., Tecchio, F. & Seri, S. (2009).
Hand somatosensory subcortical and cortical sources assessed
by functional source separation: An EEG study. Hum Brain
Mapp, 30, 660-674.
Roger, V.L., Go, A.S., Lloyd-Jones, D.M., Adams, R.J., Berry, J.D.,
Brown, T.M., Carnethon, M.R., Dai, S., de Simone, G., Ford,
E.S., Fox, C.S., Fullerton, H.J., Gillespie, C., Greenlund, K.J.,
Hailpern, S.M., Heit, J.A., Ho, P.M., Howard, V.J., Kissela,
B.M., Kittner, S.J., Lackland, D.T., Lichtman, J.H., Lisabeth,
L.D., Makuc, D.M., Marcus, G.M., Marelli, A., Matchar, D.B.,
McDermott, M.M., Meigs, J.B., Moy, C.S., Mozaffarian, D.,
Mussolino, M.E., Nichol, G., Paynter, N.P., Rosamond, W.D.,
Sorlie, P.D., Stafford, R.S., Turan, T.N., Turner, M.B., Wong,
N.D. & Wylie-Rosett, J. (2011). American Heart Association
Statistics Committee and Stroke Statistics Subcommittee. Heart
disease and stroke statistics–2011 update: A report from the
American Heart Association. Circulation, 123(4), e18-e209.
Epub 2010 Dec 15.
Rossini, P.M., Tecchio, F., Pizzella, V., Lupoi, D., Cassetta, E.,
Pasqualetti, P., Romani, G.L. & Orlacchio, A. (1998). On the
reorganization of sensory hand areas after mono-hemispheric
lesion: A functional (MEG)/anatomical (MRI) integrative
study. Brain Res, 782, 153-166.
Rossini, P.M., Tecchio, F., Pizzella, V., Lupoi, D., Cassetta, E.,
Pasqualetti, P. & Paqualetti, P. (2001). Interhemispheric dif-
ferences of sensory hand areas after monohemispheric stroke:
MEG/MRI integrative study. Neuroimage, 14, 474-485.
Rossini, P.M., Calautti, C., Pauri, F. & Baron, J.-C. (2003). Post-
stroke plastic reorganisation in the adult brain. Lancet Neurol,
2, 493-502.
Rossini, P.M., Altamura, C., Ferretti, A., Vernieri, F., Zappasodi, F.,
Caulo, M., Pizzella, V., Del Gratta, C., Romani, G.L. & Tecchio,
F. (2004). Does cerebrovascular disease affect the coupling
between neuronal activity and local haemodynamics? Brain,
127, 99-110.
Shimizu, T., Hosaki, A., Hino, T., Sato, M., Komori, T., Hirai, S. &
Rossini, P.M. (2002). Motor cortical disinhibition in the unaf-
510 G. Pellegrino et al. / Inter-hemispheric coupling changes associate
fected hemisphere after unilateral cortical stroke. Brain, 125,
1896-1907.
Siegel, M., Donner, T.H. & Engel, A.K. (2012). Spectral fingerprints
of large-scale neuronal interactions. Nat Rev Neurosci, 13, 121-
134.
Sivan, M., O’Connor, R.J., Makower, S., Levesley, M. & Bhakta,
B. (2011). Systematic review of outcome measures used in the
evaluation of robot-assisted upper limb exercise in stroke. J
Rehabil Med, 43, 181-189.
Strens, L.H.A., Asselman, P., Pogosyan, A., Loukas, C., Thompson,
A.J. & Brown, P. (2004). Corticocortical coupling in chronic
stroke: Its relevance to recovery. Neurology, 63, 475-484.
Szurhaj, W. & Derambure, P. (2006). Intracerebral study of gamma
oscillations in the human sensorimotor cortex. Prog Brain Res,
159, 297-310.
Takeuchi, N., Tada, T., Toshima, M. & Ikoma, K. (2010). Correlation
of motor function with transcallosal and intracortical inhibition
after stroke. J Rehabil Med, 42, 962-966.
Tecchio, F., Zappasodi, F., Tombini, M., Oliviero, A., Pasqualetti, P.,
Vernieri, F., Ercolani, M., Pizzella, V. & Rossini, P.M. (2006).
Brain plasticity in recovery from stroke: An MEG assessment.
Neuroimage, 32, 1326-1334.
Tecchio, F., Graziadio, S., Barbati, G., Sigismondi, R., Zappasodi,
F., Porcaro, C., Valente, G., Balsi, M. & Rossini, P.M. (2007).
Somatosensory dynamic gamma-band synchrony: A neural
code of sensorimotor dexterity. Neuroimage, 35, 185-193.
Tecchio, F., Porcaro, C., Barbati, G. & Zappasodi, F. (2007). Func-
tional source separation and hand cortical representation for
a brain-computer interface feature extraction. J Physiol, 580,
703-721.
Tecchio, F., Zappasodi, F., Tombini, M., Caulo, M., Vernieri, F. &
Rossini, P.M. (2007). Interhemispheric asymmetry of primary
hand representation and recovery after stroke: A MEG study.
Neuroimage, 36, 1057-1064.
Tecchio, F., Zappasodi, F., Porcaro, C., Barbati, G., Assenza, G.,
Salustri, C. & Rossini, P.M. (2008). High-gamma band activ-
ity of primary hand cortical areas: A sensorimotor feedback
efficiency index. Neuroimage, 40, 256-264.
Tombini, M., Zappasodi, F., Zollo, L., Pellegrino, G., Cavallo, G.,
Tecchio, F., Guglielmelli, E. & Rossini, P.M. (2009). Brain
activity preceding a 2D manual catching task. Neuroimage, 47,
1735-1746.
Traversa, R., Cicinelli, P., Pasqualetti, P., Filippi, M. & Rossini, P.M.
(1998). Follow-up of interhemispheric differences of motor
evoked potentials from the ‘affected’ and ‘unaffected’ hemi-
spheres in human stroke. Brain Res, 803, 1-8.
van Meer, M.P., van der Marel, K., Wang, K., Otte, W.M., El
Bouazati, S., Roeling, T.A., Viergever, M.A., Berkelbach van
der Sprenkel, J.W. & Dijkhuizen, R.M. (2010). Recovery of
sensorimotor function after experimental stroke correlates with
restoration of resting-state interhemispheric functional connec-
tivity. J Neurosci, 30, 3964-3972.
Volpe, B.T., Huerta, P.T., Zipse, J.L., Rykman, A., Edwards, D.,
Dipietro, L., Hogan, N. & Krebs, H.I. (2009). Robotic devices
as therapeutic and diagnostic tools for stroke recovery. Arch
Neurol, 66, 1086-1090.
Wang, L., Yu, C., Chen, H., Qin, W., He, Y., Fan, F., Zhang, Y.,
Wang, M., Li, K., Zang, Y., Woodward, T.S. & Zhu, C. (2010).
Dynamic functional reorganization of the motor execution net-
work after stroke. Brain, 133, 1224-1238.
Wheaton, L.A., Bohlhalter, S., Nolte, G., Shibasaki, H., Hattori, N.,
Fridman, E., Vorbach, S., Grafman, J. & Hallett, M. (2008).
Cortico-cortical networks in patients with ideomotor apraxia
as revealed by EEG coherence analysis. Neurosci Lett, 433,
87-92.
Wikstrom, H., Roine, R.O., Aronen, H.J., Salonen, O., Sinkkonen,
J., Ilmoniemi, R.J. & Huttunen, J. (2000). Specific changes in
somatosensory evoked magnetic fields during recovery from
sensorimotor stroke. Ann Neurol, 47, 353-360.
Zollo, L., Gallotta, E., Guglielmelli, E. & Sterzi, S. (2011). Robotic
technologies and rehabilitation: new tools for upper-limb ther-
apy and assessment in chronic stroke. Eur J Phys Rehabil Med,
47, 223-236.
Zollo, L., Rossini, L., Bravi, M., Magrone, G., Sterzi, S. &
Guglielmelli, E. (2011). Quantitative evaluation of upper-limb
motor control in robot-aided rehabilitation. Med Biol Eng Com-
put, 49, 1131-1144.
Copyright of Restorative Neurology & Neuroscience is the property of IOS Press and its content may not be
copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written
permission. However, users may print, download, or email articles for individual use.