STRUCTURE AND FUNCTION OF PARALLEL CHANNELS IN THE FOVEA

描述（由申请人提供）：我们的主要目标是研究灵长类动物视网膜向大脑传递大量信息的关键“设计”原则。它的106个轴突具有惊人的异质性。它们包括15个通道，跨越轴突直径的50倍范围和峰值速率的10倍范围。注意到中央凹回路受到空间和能量的强烈限制，我们假设存在多个通道，以便以最小的“线体积”和代谢能量成本传递信息。初步研究表明，以低信息速率（几比特/秒）传输的信道更有效（更多比特/尖峰），并且物理上更小，因此使用更少的空间和每比特可能更少的能量。因此，以不同速率获取信息的中央核（例如，膝曲M层与K层）可以以最低的成本接收信息。为了验证这一点，AIM 1将测量几种神经节细胞类型：(a)“自然信息率”（对自然图像的响应比特/峰值和比特/秒）；(b)总线材体积（体细胞+树突+轴突+终端杆）（细胞密度）；(c)相对能量成本，即线粒体含量。我们预测不同信道的自然信息速率差异很大，“低速率”信道每比特占用的空间和能量更少。

DESCRIPTION (provided by applicant): Our broad goal is to investigate key 'design' principles by which the primate retina transfers large amounts of information to the brain. Its 106 axons are strikingly heterogeneous. They comprise 15 channels, spanning a 50-fold range in axon diameter and a 10-fold range in spike rate. Noting that foveal circuits are strongly constrained for space and energy, we hypothesize that multiple channels exist in order to relay information at least cost in "wire volume' and metabolic energy. Preliminary studies suggest that channels transmitting at low information rates (few bits/second) are more efficient (more bits/spike) and physically smaller, thus using less space and probably less energy per bit. Thus central nuclei, which acquire information at different rates (e.g. geniculate M vs. K layers), can receive their messages at least cost. To test this, AIM 1 will measure for several ganglion cell types: (a) 'natural information rate' (bits/spike and bits/s in response to natural images); (b) total wire volume (soma + dendrites + axon + terminal arbor)(cell density); (c) relative energy costs, i.e., mitochondrial content. We predict that natural information rates differ strongly across channels and that 'low-rate' channels use less space and energy per bit. Noting that OFF channels are spatially finer and denser than ON channels and that 'blue/yellow' channels also occur on two spatial scales, we hypothesize that receptive field sizes and sampling rates are tuned to the distribution of information in natural scenes. To test this, AIM 2 will measure achromatic and chromatic information in natural images on scales corresponding to known receptive fields. We predict that in nature dark regions occupy higher spatial frequencies and contain more information per retinal area than bright regions and thus require finer channels with more synapses; there are analogous predictions for the blue/yellow channels. Noting that information transfer through the retina relies on 'ribbon' synapses, we hypothesize that they release and retrieve vesicles at very high rates. To test this, AIM 3 will measure release rates at the cone synapse (2-photon + electron microscopy of FM1-43 dye), and AIM 4 will measure the rates at the bipolar synapse (electrophysiology + EM of ferritin). Glaucoma, a major cause of blindness, has been attributed to both ganglion cell anoxia and reduced axonal transport. Our studies will relate ganglion cell signaling to both oxidative capacity and axonal transport capacity, and thus should extend the basic foundation for future clinical studies.