Towards Learning Retina Implants for Partial Compensation of Retinal Degenerations

摘要

The development of retina implants for blind humans suffering from various retinal degenerative diseases is the ambitious goal of several interdisciplinary research consortia worldwide. A retina encoder (RE) outside the eye has to replace the information processing of the retina. A retina stimulator (RS) implanted adjacent to the retinal ganglion cell layer at the retinal 'output', has to contact a sufficient number of retinal ganglion cells for electrical stimulation. A wireless signal- and energy transmission system (SE) has to provide the communication between RE and RS. RE consists of a group of adaptive spatiotemporal filters (RF-filters), which can be 'tuned' to various spatial and temporal receptive field (RF) properties of primate retinal ganglion cells during a learning phase. RE adaptation requires a dialog between the implant-carrying subject and the corresponding RE via a neural net.