Probing visual computations and electrical stimulation in the central macaque retina for high fidelity vision restoration

项目概要/摘要 光感受器退化导致的失明是导致残疾的主要原因。主要治疗选择 是利用视网膜前假体，直接激活视网膜神经节细胞（RGC），发送人工视觉信号 向大脑发出信号。然而，由于目前的视网膜前假体的视力恢复是有限的 粗粒度刺激无法复制自然的、光诱发的 RGC 激活模式。在自然视觉中， 约 20 种功能不同的 RGC 类型向大脑传达视觉世界的独特表征 通过协调和精确的细胞类型特异性活动模式。现代植入物不产生 高敏锐度视力的部分原因是由于粗略和非特异性，它们无法引起自然的 RGC 反应 刺激。我的研究目标是确定通过以下方法可以在多大程度上恢复中央视网膜的视力： 分析灵长类动物 RGC 对视觉和电刺激的反应。 为了实现这个目标，我首先会在猕猴的中央视网膜上进行离体实验 通过视觉和电刺激，同时使用高密度多电极阵列进行记录。识别后 主要功能不同的细胞类型，然后我们将表征它们的时空光响应特性 并确定它们与周边视网膜中相同类型的细胞有何不同。我也来比较一下效果如何 自然图像的特征由主要类型的中央和外围 RGC 表示。接下来，到 确定中央 RGC 的电激活程度，我将确定电感受野 每种细胞以及量化每种类型的 RGC 可以被选择性刺激的程度。我们随后将 测试视网膜中央主要类型的 RGC 在不激活轴突的情况下被激活的程度 捆绑。为了更好地电接触中央视网膜的 RGC，我将测试是否去除内部 限制膜降低刺激阈值并增强选择性激活。在以下情况下 单电极刺激无法选择性激活RGC，我会测试三电极是否可以 刺激可以通过电流引导更好地将电场聚焦在感兴趣的细胞上，以增强选择性 激活。最后，估计未来的植入物可以在多大程度上恢复高分辨率视力 中央视网膜，我们将通过聚合视觉和电响应特性来开发模拟 许多数据集并执行图像重建分析以量化图像的结构细节 被感知。

Project Summary/Abstract Blindness resulting from photoreceptor degeneration is a leading cause of disability. A primary treatment option is the use of epiretinal prostheses, which directly activate retinal ganglion cells (RGCs), sending artificial visual signals to the brain. However, vision restoration with current-day epiretinal prostheses is limited, due to the coarse-grained stimulation that fails to replicate natural, light-evoked RGC activation patterns. In natural vision, ~20 functionally-distinct RGC types communicate unique representations of the visual world to the brain through coordinated and precise cell type-specific patterns of activity. Modern implants don’t produce high-acuity vision in part because they fail to elicit naturalistic RGC responses, due to coarse and nonspecific stimulation. The goal of my research is to determine how well vision can be restored in the central retina by analyzing the responses of primate RGCs to visual and electrical stimulation. To accomplish this goal, I will first conduct ex vivo experiments in the central retina of the macaque monkey with visual and electrical stimulation while recording with a high-density multi-electrode array. After identifying the major functionally-distinct cell types, we will then characterize their spatiotemporal light response properties and determine how they differ from cells of the same types in the peripheral retina. I will also compare how well features from natural images are represented by central and peripheral RGCs of the major types. Next, to determine how well central RGCs can be electrically activated, I will determine the electrical receptive fields for each cell as well as quantify the extent to which RGCs of each type can be stimulated selectively. We will then test the degree to which RGCs of major types in the central retina can be activated without activating axon bundles. To achieve better electrical access to RGCs in the central retina, I will test whether removing the inner limiting membrane decreases stimulation thresholds and enhances selective activation. In cases in which selective activation of RGCs is unattainable with single-electrode stimulation, I will test whether tri-electrode stimulation can better focus the electric field on a cell of interest through current steering to enhance selective activation. Finally, to estimate how well high-resolution vision can be restored with a future implant targeting the central retina, we will develop a simulation by aggregating the visual and electrical response properties across many data sets and perform image reconstruction analyses to quantify the structural details of images that can be perceived.