Experts have high hopes for the new device. "We consider this the only current viable technology on the horizon to provide patients with high levels of cervical injury restoration and control of their limbs," says Joseph Pancrazio , director of the neural-engineering and neuroprosthesis research program at the National Institutes of Neurological Disorders and Stroke, one of the agencies funding the research.

The project is likely to be complex. Donoghue and colleagues must first make their brain chip wireless and fully implantable. (Currently, patients have some hardware protruding from their skull and are connected to a computer via wires.) An implantable system would minimize the risk of infection, and it might also help patients learn to use the system. Eberhard Fetz , a neuroscientist at the University of Washington, in Seattle, who is developing similar systems in monkeys, says that an implantable device would allow patients to use the system 24 hours a day, which would help them learn to modulate neural signals for precise control.

In the first set of tests, slated to begin next month, patients implanted with the Cyberkinetics chip will try to move a virtual arm, allowing researchers to study what level of control they could hope to achieve and to identify the muscles that need to be stimulated to elicit useful movements. Once researchers have built an implantable chip and have demonstrated that patients can sufficiently control a virtual arm, the team will start integrating the chip and the FES system.

In the long term, researchers will likely have to meld multiple devices. "To fully realize the potential of these systems, we need to think about not just a single FES system for upper limbs," says Pancrazio. "We need to think about a network of systems. The individual may need systems for ventilation, bladder control, and bowel control."