To examine the effects of using a virtual reality (VR) system to improve upper limb motor function among subacute hemiparetic stroke patients. Under this aim, the investigators have two objectives: (1) to develop and test a procedure for the quantitative assessment of stroke patients in a virtual environment; and (2) to compare the outcome measures of the two groups of stroke patients (VR-intervention vs. conventional-control) on these two modes of training.

After being informed about the study and potential risks, all participants who gave written consent will undergo the study. A randomized controlled trial with pre-/post-study, of 42 stroke patients with upper limb weakness (21 patients in the control group and 21 patients in the intervention group), 18-65 years old, all are newly admitted stroke patients in the rehabilitation clinic from one hospital. Participants in the intervention group will receive 60 minutes of VR training to improve upper limb function and daily living activities. Participants in the control group will receive 60 minutes of conventional training to improve upper limb function and daily living activities. The training includes grasp, grip, pinch, gross, and finger-hand coordination movement. The training will perform 5 times/week for 4 weeks. Two outcome measures: (a) The Barthel Index and (b) The Fugl-Meyer Assessment, time spent to complete a task were used to compare two groups.

Inclusion Criteria:

1. New in-patients clinically diagnosed hemiparetic stroke patients;
2. Patients with a history of the first incidence of stroke with upper limb weakness with a motor impairment upper limbs as a primary deficit;
3. muscle strength > 2/5 on the Medical Research Council (MRC) scale;
4. good cognitive ability as indicated by a score > 15 on the Mini-Mental State Examination (MMSE); Motricity Index sore > 26; and
5. ability to communicate and sign the consent.

Exclusion Criteria:

1. In-patient with ataxia or any other cerebellar symptoms;
2. orthopedic alterations or pain syndrome of the upper limb;
3. peripheral nerve damage affecting the upper extremities;
4. visual or hearing impairment (does not allow the possibility of interaction with the VR system);
5. severe hemispatial neglect;
6. severe shoulder or hand pain (unable to volitionally position the hand in the workspace without pain);
7. insensate forearm and /or hand, edema of the affected forearm and /or hand;
8. uncontrolled seizures disorder;
9. severe depression (>13 on beck Depression inventory fast screen); and
10. severely impaired cognition or comprehension.

The assessor will assess participants' outcomes after they complete the rehabilitation training. The assessor does not know who is undergoing the VR (intervention) or non-VR (control) rehabilitation training