The Drosophila eye consists of ∼750 single units called ommatidia, which contain eight photoreceptor neurons (R cells) each. R-cell axons project to two visual processing centers in the brain, the lamina and the medulla. The R1–R6 cells connect with lamina neurons L1–L5 (superposition, which will be discussed latter). R7 and R8, however, bypass the lamina and innervate the second optic ganglion, the medulla, targeting two distinct synaptic layers (Fig. 2A). Visual space within the medulla (processing center, has 800 column and recieve info directly from retina or indirectly from lamina) is represented in a topographically ordered array of neuronal columns corresponding to the number of ommatidia in the retina. Every column consists of neurites of ∼60 different cell types and receives sensory input from the retina via R7 and R8, and indirectly through L1–L5 from the lamina. Based on a precise wiring pattern, visual information from one point in space is relayed to only a single column in the medulla.
Every thing begins with the fact that, among 8 phtorecptor neurons (R1-R8) in one ommatidium, each has a slightly direction shift, which means if we take all the #3 fibers from each ommatidum, they will point at different orientation and meaningless for the visual processing. Therefore, smart drosophlia find, if I pick #3 fiber from the first ommadtidum, #4 from the second ommatidium and #5 from the third ommaditium those three fibers will point at the same direction and if we collect them at the next level (lamina) (btw, for fly visual system, retina is the outtermost part have 800 ommitidum each contain 6 photorecptor in 6 directions, then they will be sorted into 800 bundles but each with the same direction of 6 neurons at lamina, a thin layer between retina and medula, then medula the information processing part, then lobula), information will be relayed and processed more conviently since they have the same direction info.
This process occurs in two steps: R1–R6 axons from a single ommatidium initially bundle together forming an axonal fascicle and project to the developing lamina neuropil. In a subsequent developmental phase, individual R cell growth cones extend laterally out of their fascicle towards a neighboring bundle of lamina neurons to form a lamina cartridge.
At the thin layer of lamina, 6 retina with sorted the same direction enter the bundle in lamina, these bundle have the name of cartridge (or neuronommatidum). Within the cartridge bundle, 6 afferent fibers (R1-R6 from retina, but from 6 different ommatidums, rememebr the cyclic twist?) and 3 efferent fibers (L1-L3, know we know there are 5, L1-L5) interact and only interact within the cartridge. Then L1-L3 receive the info the go further, whereas R1-R6 finish their roles and disapper abrutly.
Getting pieces of evidences together, the key conclusion is: EACH cartridge only receive the visual information from one position from environment only. The optical resolution in spatial is only decided by the number of cartridge and irrevelent with that of retina photorecepter neurons. The latter may serve another funciton in visual resolution such as selective sensitivity to a specific specturm of light wave leangth. This idea is summarized as the figure below.
In a certain cartridge, all the six R1-R6 are the neurons whose axis is parallel to the line of sight of that element.
