A Retina-Inspired Computational Model for Stimulation Efficacy Characterization and Implementation Optimization of Implantable Optogenetic Epi-Retinal Neuro- Stimulators

In this paper, a biologically-informed computational framework is developed to model the efficacy and to optimize the implementation of an implantable epi-retinal prosthesis that performs optogenetic stimulation through a $\mu $ LED array. The developed model is capable of translating visual stimulus inputs into corresponding signals evoked in the transfected retinal cells through optogenetic stimulation, calculating the subsequent neuronal activities of the following retinal layers, and estimating the resulted brain’s visual perception. As such, it can model and quantitatively analyze the impact of optical stimulation parameters (intensity, frequency, directivity, wavelength, etc.) and the $\mu $ LED array’s physical specifications (array size, density, pitch, implantation location, etc.) on the efficacy of the stimulation. Using this model, we compared optical and electrical stimulations in terms of the structural similarity between their induced visual perception in the brain and the visual stimulus input. We showed that thanks to the cell-type specificity of optogenetic stimulation, it can induce more relevant visual perception qualities than electrical stimulation. We also showed that its resulted visual perception substantially improves with scaling the stimulator’s array size. The model was also used to qualitatively and quantitatively analyze the impact of parameters such as implantation location, light intensity, single- and dual-wavelength stimulation, and illumination divergence angle on the quality of the optical-stimulation-induced visual perception. In each case, the simulation results were followed by our interpretation from a biological point of view. More importantly, in each case, we discussed how the results could be used for optimizing different parameters of an implantable optogenetic stimulator to achieve maximum efficacy and energy efficiency.