Learning retina implants with epiretinal contacts

Area of Science:

• Biomedical Engineering
• Neuroscience
• Ophthalmology

Background:

• Retinal degenerative diseases cause blindness, necessitating advanced visual prostheses.
• Current research focuses on developing retina implants to restore partial or functional vision.
• Interdisciplinary consortia are exploring various approaches to visual restoration.

Purpose of the Study:

• To develop a novel retina implant system for restoring vision in blind individuals.
• To achieve moderate vision, including object location and shape perception, in the initial stage.
• To progress towards reading quality in subsequent stages of development.

Main Methods:

• Designing a retina implant with an external retina encoder (RE) and an internal retina stimulator (RS).
• Utilizing a wireless system for signal and energy transmission between the RE and RS.
• Implementing a learning RE with tunable spatiotemporal receptive field (RF) filters to map visual patterns onto spike trains.

Main Results:

• The learning RE successfully approximates primate retinal ganglion cell receptive field properties.
• A concept for unsupervised learning trains individual RF filters through human subject interaction.
• The system maps visual patterns onto spike trains for stimulated retinal ganglion cells.

Conclusions:

• The developed retina implant system shows promise for restoring vision in patients with retinal degenerative diseases.
• The learning RE, through adaptable RF filters, aims to optimize visual perception for central visual system integration.
• This approach represents a significant step towards functional vision restoration using visual prostheses.