This prosthesis bypasses the damaged retinal cells and artificially stimulates the optic nerve directly through implantation of an electrode around the nerve that automatically expands and contracts with nerve movements. The device processes images captured by a head-mounted mini-ature camera and transmits them to the optic nerve, where the patient can extract visual information from the images, explained Dr. Veraart, who was the biomedical engineering coordinator for the project and has now retired.

The first patient, who had no residual vision, received an intracranially implanted (IC) self-sizing spiral cuff electrode with four contacts around her right optic nerve in 1998. In 2000, the percutaneous lead was replaced by an implanted stimulator and antenna for telemetry. The second patient received an intraorbitally implanted (IO) spiral electrode with eight contacts around his right optic nerve in 2004.

Recognition of patterns

"With the same patient we have also observed good object localization, discrimination, and grasping, but again with relatively long delays," he added.

These tasks were typically completed in about 1 minute.

The outcome of these experiments has shown that a relatively basic algorithm can convey information about the visual scene efficiently, Dr. Veraart said. Investigators also observed that performance improves as the number of phosphenes used in the algorithm increases but that a longer training time is required to accomplish the tasks in the experiment. More training was required to accomplish the localization and discrimination tasks than for grasping objects.

Dr. Veraart also noted evidence of good biocompatibility with the prosthesis. The second implantation took place 6 years after the first, with several modifications.

"Because intracranial implantation was very invasive, we tried to design a new surgical method of intraorbitally implantation," Dr. Veraart said. This resulted in only 2 days of hospitalization com- pared with 9 for the first case.

The safer and dramatically less invasive medial transconjunctival approach will be used in future implantations despite the fact that the first patient on whom this technique was used was unable to achieve closed-loop stimulations because of large, diffuse phos-phenes, Dr. Veraart said.

A multidisciplinary team of collaborators from several European countries has been involved in this research, including electronics laboratories in France and Belgium, firms in Germany, and the Saint Luc University Hospital in Belgium. Jean Delbeke, MD, PhD, can be contacted for further information.