To investigate whether early therapy with a novel robotic device can reduce motor impairment and enhance functional recovery of poststroke patients with hemiparetic and hemiplegic upper limb.A single-blind randomized controlled trial, with an 8-month follow-up.Neurologic department and rehabilitation hospital.Thirty-five patients with acute (< or =1 wk of onset), unilateral, ischemic embolic, or thrombotic stroke.Patients of both groups received the same dose and length per day of standard poststroke multidisciplinary rehabilitation. Patients were randomly assigned to 2 groups. The experimental group (n=17) received additional early sensorimotor robotic training, 4 hours a week for 5 weeks; the control group (n=18) was exposed to the robotic device, 30 minutes a week, twice a week, but the exercises were performed with the unimpaired upper limb. Training by robot consisted of peripheral manipulation of the shoulder and elbow of the impaired limb, correlated with visual stimuli.The Fugl-Meyer Assessment (FMA) of upper-extremity function (shoulder/elbow and coordination and wrist/hand subsections) to measure each trained limb segment; the Medical Research Council (MRC) score to measure the strength of muscle force during 3 actions: shoulder abduction (MRC deltoid), elbow flexion (MRC biceps), and wrist flexion (MRC wrist flexors); the FIM instrument and its motor component; and the Trunk Control Test (TCT) and Modified Ashworth Scale (MAS).Compared with the patients in the control group, the experimental group showed significant gains in motor impairment and functional recovery of the upper limb after robot therapy, as measured by the MRC deltoid (P< or =.05) and biceps (P<.05) scores, the FMA for the proximal upper arm (P<.05), the FIM instrument (P<.05), and the FIM motor score (P<.01); these gains were also sustained at the 3- and 8-month follow-up. The FMA and MRC wrist flexor test findings did not differ statistically either at the end of training or at the follow-up sessions. We found no significant differences in MAS and TCT in either group in any of the evaluations. No adverse effects occurred and the robotic approach was very well accepted.Patients who received robotic therapy in addition to conventional therapy showed greater reductions in motor impairment and improvements in functional abilities. Robotic therapy may therefore effectively complement standard rehabilitation from the start, by providing therapeutic support for patients with poststroke plegic and paretic upper limb.