Automatic Identification of Axon Bundle Activation for Epiretinal Prosthesis
Summary
This study introduces an algorithm to detect axon bundle activation in retinal prostheses, improving visual restoration. The method accurately identifies stimulation thresholds, reducing uncontrolled visual percepts for better artificial vision.
Area of Science:
- Biomedical Engineering
- Neuroscience
- Ophthalmology
Background:
- Restoring high-fidelity vision with retinal prostheses requires selective cell activation.
- Inadvertent activation of distant retinal ganglion cells (RGCs) via axon bundles limits artificial vision quality.
Purpose of the Study:
- To develop an algorithmic solution for detecting axon bundle activation in bi-directional epiretinal prostheses.
- To overcome challenges in identifying axonal stimulation of RGCs with unknown locations.
Main Methods:
- An algorithm was developed to detect axon bundle activation using electrical recordings.
- It determines stimulation current amplitudes that trigger axonal activation.
- The method leverages spatiotemporal spike characteristics to detect small axonal spikes.
Main Results:
- The algorithm was validated using ex vivo macaque retina experiments.
- It achieved a ±10% accuracy for bundle activation thresholds compared to manual identification in 88% of electrodes.
- A high correlation coefficient of 0.95 was observed.
Conclusions:
- A simple, accurate, and efficient algorithm for detecting axon bundle activation in epiretinal prostheses is presented.
- This algorithm can enhance closed-loop control in future prostheses.
- The method shows broad applicability for other neural implants.
