A Patient-Specific Computational Framework for the Argus II Implant.

Goal: Retinal prosthesis performance is limited by the variability of elicited phosphenes. The stimulating electrode's position with respect to retinal ganglion cells (RGCs) affects both perceptual threshold and phosphene shape. We created a modeling framework incorporating patient-specific anatomy and electrode location to investigate RGC activation and predict inter-electrode differences for one Argus II user.
Methods: We used ocular imaging to build a three-dimensional finite element model characterizing retinal morphology and implant placement. To predict the neural response to stimulation, we coupled electric fields with multi-compartment cable models of RGCs. We evaluated our model predictions by comparing them to patient-reported perceptual threshold measurements.
Results: Our model was validated by the ability to replicate clinical impedance and threshold values, along with known neurophysiological trends. Inter-electrode threshold differences in silico correlated with in vivo results.
Conclusions: We developed a patient-specific retinal stimulation framework to quantitatively predict RGC activation and better explain phosphene variations.