Bipolar
A @bipolarLayer class is a subclass of the general @cellLayer class. We expect that a number of different neural structures will be defined as layers (bipolar, rgc, LGN, V1) that contain mosaics of cells. The bipolar layer defines organizes a collection of @bipolarMosaic objects. The bipolar layer has a visualization method (@bipolarLayer.window). Over time, we expect the layer information and methods will increase.
The @bipolarMosaic represents a collection of bipolar cells that span a portion of the visual field. Each mosaic has a method (@bipolarMosaic.compute) that converts the input (@coneMosaic.current) into the bipolar mosaic output. Each bipolar mosaic stored in the layer has its own spatial and temporal parameters. They can also implement different computations. At this time, we are experimenting with models that implement rectification.
ISETBIO includes five pre-defined bipolar mosaic names: on and off diffuse, on and off midget, and onsbc (on small bistratified cell). These bipolar types have preset parameters that reflect the general properties of these types of bipolars (larger spatial RF sizes, different cone type contributions). But the parameters can always be adjusted to reflect differences between species, position in the visual field, and health or disease.
In which we talk about cone mosaic frame distances and then samples on the input layer.
Receptive field size is chosen to incorporate about 2x2 cones.
The temporal sampling rate parameter
The bipolar temporal impulse response (IR) has not been established in the literature. To set the impulse response we accepted the temporal impulse response of the RGC from the literature, and made the combination of the cone photocurrent and the bipolar IR equal to the measured RGC impulse response. For this reason, the RGC (described elsewhere) impulse response is assigned a delta function.
The bipolar object also allows for the simulation of nonlinear subunits within retinal ganglion cell spatial receptive fields.
This page goes into the details of how we chose to implement the default parameters of the bipolar model. These parameters can be adjusted by the user as thinking evolves. Our defaults are based on the papers described in the [bipolar reference page](bipolar reference).
