The two cerebral hemispheres find themselves in a state of balanced mutual inhibition. A stroke with involvement of motor function leads to a reduced excitability in affected hemisphere M1 and to an increased excitability of contralateral M1. Stroke therefore might impair interhemispheric balance, leading to a decreased inhibition of contralesional hemisphere by ipsilesional hemisphere and, in turn, to an increased inhibition of ipsilesional hemisphere by contralesional hemisphere. Permanence of healthy hemisphere hyperactivation in chronic phase after stroke is usually index of little functional recovery and is correlated with a greater ipsilateral structural damage.

Robot-mediated physical therapy is an innovative rehabilitation technique that is effective in stroke patients. In this study, the investigators will add a non-invasive brain stimulation protocol with Transcranial stimulation with direct current (tDCS) to the robotic treatment in chronic stroke patients. tDCS is a non-invasive brain stimulation technique that is able to modulate cortical excitability. The hypothesis is that dual t-DCS (ipsilesional excitation and concomitant contralesional inhibition) could restore interhemispheric balance improving the benefits of robotic therapy with Armeo Power®.

The general theory of the inhibitory interhemispheric competition speculates that, under physiological conditions, the two cerebral hemispheres find themselves in a state of balanced mutual inhibition. Accordingly, some electrophysiological and functional imaging studies have shown that a stroke with involvement of motor function leads to a reduced excitability in affected hemisphere M1 and to an increased excitability of contralateral M1. Stroke therefore might impair interhemispheric balance, leading to a decreased inhibition of contralesional hemisphere by ipsilesional hemisphere and, in turn, to an increased inhibition of ipsilesional hemisphere by contralesional hemisphere.

This imbalance in cerebral hemispheres excitability tends to decline over the functional recovery in the first months after stroke, in proportion to the functional recovery of the ipsilateral structures.

Alternatively, permanence of healthy hemisphere hyperactivation in chronic phase after stroke is usually index of little functional recovery and is correlated with a greater structural damage to the ipsilesional corticospinal tract and transcallosal fibers, although some studies have suggested a vicarious role of healthy hemisphere.

Recently, some authors have proposed a structural reserve-based bimodal model of recovery, for which the vicarious model would be valid only for patients with poor structural reserves (and therefore more extensive damage), while the interhemispheric competition model would be valid for patients with larger structural reserves.

In the light of these data, it is conceivable that procedures aiming at promoting the re-establishment of a normal balance in interhemispheric interaction can facilitate motor function recovery of the paretic limb, at least in patients with appropriate structural reserves.

Transcranial stimulation with direct current (tDCS) is a non-invasive brain stimulation technique that uses low intensity electrical current at constant voltage and variable duration that is able to modulate cortical excitability efficiently, lastingly (with effects that begin during stimulation and that last for a few hours from the end of the stimulation) and safely in healthy subjects: moreover, it is able to produce changes of excitability in both directions (i.e. towards an increase and towards its reduction) in dependence of electrodes positioning on the scalp.

Specifically, t-DCS appears to induce metaplasticity phenomena, that is a polarity-dependent modulation of the response to subsequent synaptic potentiation protocols. Considering these results and the conclusions of a recent meta-analysis, it seems appropriate to expose a patient to t-DCS immediately before or in the early stage of intervention aiming at promoting synaptic plasticity, such as neuro-rehabilitation techniques.

Alongside the traditional physical therapy techniques, during the last twenty years many research groups around the world have started to develop robots supporting physical treatments.

A review on robot-mediated physical therapy clinical trials shows how this innovative rehabilitation technique is effective both in acute and chronic stroke patients, although it is not proven its superiority over traditional rehabilitation methods.

In this study, the investigators will use a robotic exoskeleton that allows targeted interventions on specific tasks in the three spatial dimensions; this robot, marketed under the name Armeo Power® (Hocoma AG, Switzerland), is one of the most advanced rehabilitation devices for the upper limb. The effectiveness of this tool has been demonstrated in a recent multicenter trial of chronic stroke patients with moderate-severe upper limb motor deficit.

The purpose of this study is to evaluate the benefits arising from the combination of dual t-DCS (ipsilesional excitation and contralesional inhibition) and robotic therapy with Armeo Power® in the treatment of chronic stroke patients with upper limb paralysis.

This study is a double blind randomized controlled trial in which chronic stroke patients with upper limb paresis undergone robotic treatment with Armeo Power® will be randomized to concomitant dual tDCS (excitatory on injured side and inhibitory on health side) o sham tDCS.

All patients will undergo an initial evaluation before starting treatment (between 30 and 15 days before the start) where it will be filled in a form with the medical history and the personal data and will be signed informed consent.

Once eligible, the patient will be subjected to a first clinical evaluation using validated scales, a kinematic evaluation of upper limb motor performance with ArmeoPower® and a neurophysiological evaluation.

At the end of the evaluation, every patient will be assigned through a randomized system to one of two study groups:

• Group A: robotic treatment with Armeo Power® associated with REAL dual tDCS (anodic on injured hemisphere, cathodic on healthy hemisphere) - 40 patients
• Group B: robotic treatment with Armeo Power® associated with SHAM dual tDCS - 40 patients

Immediately prior the start of treatment (time T-1), every participant will perform a second clinical assessment and kinematic evaluation and will begin treatment on the same day in the gym of the Private Department of Saint Lucia Foundation.

The ArmeoPower® device is an exoskeleton robot for upper limb rehabilitation able to allow the patients' execution of movements of proximal (shoulder-elbow movements) and distal (movements of the wrist) segments of upper limb.

Once the arms of the robot are adapted to patent's upper limb, the exoskeleton is mounted and set for defining the action areas and the volumes of exploration.

At the beginning of each session, ArmeoPower® exoskeleton will be mounted according to the patients' stored settings, which takes about 10 minutes.

Once mounted, it will proceed to real or sham tDCS stimulation, depending on the group the patient belong to.

Participants in group A will perform real dual tDCS stimulation immediately before each session of the robotic treatment, while patients in group B will perform the treatment with robotics platform with the same intensity and duration of group A but preceded by sham dual tDCS stimulation.

Real dual tDCS stimulation requires the application of the anode over the injured motor cortex (M1) and of the cathode on the healthy motor cortex (M1), so as to simultaneously perform an excitability suppression of healthy hemisphere and an activation of the ischemic cortex.

Motor cortex will be simultaneously stimulated at an intensity equal to 1-2 mA and for a maximum duration of 20 minutes. Sham dual t-DCS requires the application of the electrodes in the same positions and for the same duration, but without the delivery of electrical stimulation.

The robotic rehabilitation treatment will be started immediately after the end of stimulation, and the actual treatment will last for 30 minutes. The robot is able to relieve the weight of the arm and assisting the patient during motor execution. The patient sees on a monitor the targets upon which he should move the effector and executes the movement through the active assistance of the robot, which compensates the patient's motor deficits. Videogame-based exercises and functional training exercises of daily life activities are carried out with visual, acoustic and performance feedback. A physical therapist with a specific training certificate is always present during the sessions, with 1:1 ratio with patients.

Each patient will perform 5 sessions per week for two consecutive weeks, after which will be performed clinical, kinematic and neurophysiological evaluations expected to T0; the patient will then be re-evaluated in follow-up at one month (T1) and at 3 months after the end of treatment (T2), with clinical and kinematic assessments.

Data will be analyzed using parametric statistics if will appear to be normally distributed (based on Kolgomorov-Smirnov test with Lilliefors correction), in particular by applying a two-way mixed model analysis of variance (with an intra- and inter-factor between subjects). If the data won't appear to be normally distributed, they will be analyzed using Friedman's analysis for the intra-subject variations, and by Mann-Whitney u-test for variations between the two groups. The level of significance will be set at 0.05 for all analysis, except for post-hoc to which is applied the Bonferroni correction.

Guidelines for the safe use of non-invasive brain stimulation (Wassermann, 1998) will be respected; the participants should be able to signed an informed consent.

There are not literature report of adverse events or contraindications with regard to the use of ArmeoPower® robotics platform, except for patients with severe deformities of the paretic upper limb, that are excluded from this study. Data on patients will be anonymous and stored with password protection. There will be no samples of biological material, drug delivery, healthy volunteers or animals.

Inclusion Criteria:

• Subjects with isolated ischemic stroke, confirmed by brain imaging (MRI, CT)
• subjects who retain sufficient cognitive functions and language to follow the instructions given by doctors and therapists
• subjects with basal Modified Ashworth Scale score under 3
• subjects with basal Fugl-Meyer score ≥ 3 so that they are not completely paralyzed
• subjects who have signed informed consent to participate in this study
• subjects that show stable conditions in the two pre-treatment evaluations, in order to avoid "Hawthorne effect"

Exclusion Criteria:

• subjects with chronic paretic limb deformities
• subjects with complete and flaccid paralysis of all motor performances of shoulder and elbow;
• subjects with severe hemineglect (Pizzamiglio battery for unilateral spatial neglect including letter cancellation test, barrage tests, reading and Wundt-Jastro area illusion test; patients diagnosed with neglect if 3 of 4 items of this battery are present)
• subjects showing an increase in the Fugl-Meyer more than 2.1 points in the second pre-treatment clinical evaluation, compared to the score of the first evaluation
• subjects with contraindications to the execution of transcranial magnetic stimulation - TMS - (pacemakers, metal implants)
• subjects with epilepsy
• Previous neurosurgical interventions
• Severe upper limb osteoporosis
• Upper limb strength or joint movement limitation due to previous fractures
• Upper limb strength or joint movement limitation due to previous surgical interventions
• Mini Mental State Evaluation (MMSE) <24