54 inpatients participants were randomly divided into two groups (experimental and conventional). Individual of experimental group were treated according to an established protocol for ARMEO Spring (30 minutes/session with "Armeo Spring" and 30 minutes/session with conventional treatment 5 days/week for 6 weeks). The conventional treatment consists of passive and active assisted mobilization of the upper limbs, traditional training based on the Bobath concept. Inpatients of control group were treated with conventional treatment with training session of 60 minutes 5 days/week for 6 weeks. All patients were evaluated by a blinded observer using the outcomes tests at enrollment (T0), after the treatment (T1) and at follow up 6 weeks later (T2). We assessed the impact on functional recovery (Functional Independence Measure - FIM scale), strength (ARM Motricity Index-MI), spasticity (Modified Ashworth Scale-MAS) and pain (Numeric Rating Pain Scale -NRPS).

Authors conducted a double blind randomized controlled trial. Informed consent was obtained from all participants and procedures were conducted according to the Declaration of Helsinki. The protocol (N° U0074917/11110) was approved by the Local Ethical Committee of Bergamo, Italy. A number of clinical trials have shown significant advances in upper limb recovery with the use of different sensory - motor techniques, including intensive repetitive movement, constraint-induced movement therapy, functional electrical stimulation treatment, the use of robot-assisted therapy in association with virtual reality. Robot-assisted virtual reality intervention has been shown more effective than conventional interventions and achieved more improvement in upper limb function, however the effect size reported by recent reviews was small and this datum is always reported as a cost-benefit ratio to challenge the use of virtual reality technology in rehabilitation. Robot aided rehabilitation is increasingly used in stroke rehabilitation, with a broad spectrum of applied technology from motor to non-motor aided systems, posing the problem to match the clinical need of the patient with the proper device. In the immediate post stroke period the motor recovery usually do not allow the use of non-motored robot device while the use of motored robot aid after months can be too delayed to obtain some valuable clinical results. The wide range of available devices are certainly a richness in the clinical possibility but also a critical factor in selecting a suitable technology tailored for the clinical feature of the patient at the actual state of the art. This can affect the comparison and interpretation of the literature so far published. Virtual reality therapy recreates favorable conditions to motor learning. Functional recovery is achieved through use-dependent cortical reorganization. The time/intensity of its application is therefore a pivotal point in this learning process. Its duration is not standardized and can varies from 3-4 to 20 hours of total treatment making results accordingly variable, adding further bias in data interpretation. A prerequisite to gain the best results is patients selection and early application when is possible.All these factors, have been ascribed as possible causes of small effect size reported in recent literature in comparing robot-aided virtual reality rehabilitation versus traditional rehabilitation alone.In a group of patients with hemiparesis following stroke, we compared the efficacy of a neurorehabilitation program consist of combination of non-motor robot-assisted therapy with virtual reality (Armeo®Spring) to conventional therapy with the primary aim to verify if the punctual application of what suggested by the single papers is able to improve differences of the efficacy between treatments and, therefore, creating a better effect size. And, as a secondary arm, if it is possible to improve the clinical picture also in post-acute stroke patients and as a secondary aim, if motor selection and intensive treatments can improve the small effect size reported by the literature.

Inclusion Criteria:

• The clinical diagnosis of stroke
• post stroke hemiparesis
• maximum six months from stroke
• stability of the clinical picture at the time of roll-in
• minimum level of upper arm motility (>2) movement against gravity
• trunk control and ability to maintain sitting position for at least 120 minutes

Exclusion Criteria:

• Hemiplegia of other aetiology ( neurodegenerative, neoplastic)
• Presence of articular ankyloses, spasms and/or severe muscle spasticity with complete loss of movement in any of the three major joints
• Instability of upper limb bone (not consolidated fractures)
• Presence of cognitive impairment (MMSE<=21) and/or psychiatric disease
• Concomitant disease that could prevent the rehabilitation program (respiratory failure, heart failure, osteomyelitis, thrombophlebitis and other clinical condition that are against rehabilitation treatment)
• Ulcer sores that can contraindicate the use to ARMEO Spring
• Ashworth > 3 (for each of the three upper limb joints)