If you are running the simulation of color selective cells (Simulation menu ) you can play with color selective cells. Right click in the tangent screen to pick a stimulus and background color .
By choosing a cell using the Cell menu, you can monitor different color-selective cone types (L, M, or S) and different types of color-selective horizontal cells, bipolar cells, and ganglion cells .
Inspecting this list shows a lot of replication over different colors. One thing that may stand out, however, is the lack of a blue-OFF channel. No one has seen evidence of a channel to represent blue-OFF. Other than the blue channel, all the channels represent color information in the usual rectified manner: the ON and OFF channels separately transmit information about local contrast that is brighter or darker than the surrounding context.
The 3 types of cones model the LMS (Long, Middle, Short) color selectivities of biological cones. You can play with these photoreceptor models by varying the stimulus color and observing the LMS values of the colors you choose. The cone spectral responses are shown in these curves:
Note how the L and M senstivities are very similar. The following plot shows the difference between each curve and the average of L and M:
The L and M horizontal cells average over the L and M cones respectively.
The L+M and L+M+S horizontal cells average over multiple types of cones and represent a measure of the luminance. Since there are few S cones compared with L and M cones, these models are pretty close to each other in reality. In the simulation, they differ because the simulation has equal numbers of LM and S cones.
The most realistic simple model of color-selective horizontal cells is that they don't exist; there is no strong evidence for a color-selective invervation of photoreceptor cone pedicles by horizontal cell processes. Any color-selectivity in horizontal cells appears to be accidental and due to random perturbations in the cone types that are innervated by the horizontal cells. Since most cones are either L or M type, most horizontal cells will be of type L+M. This L+M cell is the horizontal cell type that is used for all the bipolar cell modeling. It is most responsive to L+M= yellow.
The simulations of color-selective bipolar cells are closest to midget bipolar cells, which connect directly to a single cone. These cells appear to have a "private line" to the brain, a term coined by one of the pioneers of work on the color system.
Each of the simulated color-selective bipolar cells is connected to a single cone and with the opposite type of synapse to the L+M horizontal cell.
The receptive fields look like the following:
For example, the Bipolar red-on cell is excited by the center L cone and inhibited by the L+M horizontal cell..
What about the Bipolor red-off cell? It is excited by darkness in the center and (L+M= yellow) in the surround.
The best stimulus for each cell type is shown in the following:
The color selective ganglion cells.are simply connected directly to the corresponding bipolar cells. They have exactly the same receptive fields.