The patients underwent forty 1h Armeo-P training sessions (i.e. five times a week for eight consecutive weeks). During the first session, the device was adjusted to the patient's arm size and the angle of suspension. The working space and the exercises were selected once the UL had been fitted with the system. All the subjects in the arm received a focal belly-muscle vibration on the spastic antagonist muscles (i.e. triceps brachialis-TB, deltoid-DE, and supraspinatus-SS) during shoulder abduction and elbow extension. MV was delivered by a pneumatic vibrator powered by compressed air, wired to appropriate-muscle probe diameter (up to 2cm2). MV was set at a frequency of 80Hz and an individually adjusted vibration amplitude so that it was just below the threshold for perceiving an illusory movement. The investigators chose such set up to avoid any signs of muscle contraction potentially reflecting either possible voluntary movement or occurrence of the tonic vibration reflex (TVR).

The patients underwent the same Armeo-P trainingas the experimental group. Only the vibration protocol was didderent. Indeed, in the control group a sham vibration was used, while in the experimental group, patients underwent a real one.

Sham vibration was delivered to the control group using the same procedure of the experimental group;however, vibration intensity was subthreshold (i.e. 50mBar below the threshold).

All the subjects assigned to the experimental received a focal belly-muscle vibration on the spastic antagonist muscles (i.e. triceps brachialis-TB, deltoid-DE, and supraspinatus-SS) during shoulder abduction and elbow extension. MV was delivered by a pneumatic vibrator powered by compressed air (Vibraplus; @Circle, San Pietro in Casale, Italy) (fig. 2) wired to appropriate-muscle probe diameter (up to 2cm2). MV was set at a frequency of 80Hz and an individually adjusted vibration amplitude (approximately 2mm in indentation depth, mean pressure of 250mBar, which effectively induce inhibition of the monosynaptic reflex) so that it was just below the threshold for perceiving an illusory movement. We chose such set up to avoid any signs of muscle contraction potentially reflecting either possible voluntary movement or occurrence of the tonic vibration reflex (TVR).

The primary outcomes consisted in the effects of MV and Armeo Power® on spasticity by a clinical (MAS from spastic agonist muscles causing arm adduction, inward rotation, and flexion, and forearm flexion)

The investigators measured the effects of MV and Armeo Power® on UL motor function impairment (measured by Fugl-Meyer Assessment of UL motor recovery after stroke -FMA).