OBJECTIVE To investigate the feasibility of implanting microstimulators to deliver programmed nerve stimulation for sequenced muscle activation to recover arm-hand functions. DESIGN By using a minimally invasive procedure and local anesthesia, 5 to 7 microstimulators can be safely and comfortably implanted adjacent to targeted radial nerve branches in the arm and forearm of 7 subjects with poststroke paresis. The microstimulators' position should remain stable with no tissue infection and can be programmed to produce effective personalized functional muscle activity with no discomfort for a preliminary 12-week study. Clinical testing, before and after the study, is reported in the accompanying study. SETTING Microstimulator implantations in a sterile operating room. PARTICIPANTS Seven adults, with poststroke hemiparesis of 12 months or more. INTERVENTION Under local anesthesia, a stimulating probe was inserted to identify radial nerve branches. Microstimulators were inserted by using an introducer and were retrievable for 6 days by attached suture. Each device was powered via a radiofrequency link from 2 external cuff coils connected to a control unit. MAIN OUTCOME MEASURES To achieve low threshold values at the target sites with minimal implant discomfort. Microstimulators and external equipment were monitored over 12 weeks of exercise. RESULTS Seven subjects were implanted with 41 microstimulators, 5 to 7 per subject, taking 3.5 to 6 hours. Implantation pain levels were 20% more than anticipated. No infections or microstimulator failures occurred. Mean nerve thresholds ranged between 4.0 to 7.7 microcoulomb/cm(2)/phase over 90 days, indicating that cathodes were within 2 to 4 mm of target sites. In 1 subject, 2 additional microstimulators were inserted. CONCLUSIONS Microstimulators were safely implanted with no infection or failure. The system was reliable and programmed effectively to perform exercises at home for functional restoration.