Cortical neural prostheses for vision restoration

In this thesis, we demonstrate that the primary visual cortex (V1) can serve as a viable entry point for a high-density neural prosthesis, capable of bypassing the eyes to restore "phosphene" vision. By implanting a record-breaking 1,024 electrodes in macaque monkeys, we proved that microstimulation can successfully evoke complex phosphene patterns, allowing subjects to recognize shapes and letters with high spatial resolution. To ensure these systems can be calibrated efficiently, we developed the "NEUmap" algorithm, which uses spontaneous brain activity to map the visual field without requiring days of active patient feedback. We also addressed the technical hurdle of "closed-loop" communication by optimizing artifact-removal techniques, ensuring that we can simultaneously stimulate and record neural responses. However, our long-term findings highlight a critical "biological wall": over three years, the chronic inflammatory response and fibrotic encapsulation led to a significant loss of functional electrodes. Our work concludes that while we have mastered the "code" for artificial vision, the next frontier must focus on developing flexible, biocompatible materials that can withstand the brain’s immune response to maintain a stable interface for decades