-
Notifications
You must be signed in to change notification settings - Fork 42
Retinal ganglion cells
The retinal ganglion cell layer (@rgcLayer) is a subclass of the cell layer (@cellLayer) class. The rgc layer organizes a collection of @rgcMosaic objects. The rgc layer object stores properties that are common to all of the rgc mosaics. The rgc layer has a visualization method (@rgcLayer.window).
The rgcMosaic object represents a mosaic of RGCs of a specific cell type. The rgcMosaic object takes one or more bipolar objects as input and, using the compute method (@rgcMosaic.compute), converts these inputs into the linear and spiking responses of each cell in the rgc mosaic.
@rgcMosaic
The user may choose from several computational models:
-
The linear-nonlinear-Poisson (LNP) model: see Chichilnisky and Kalmar, "Functional asymmetries in ON and OFF ganglion cells of primate retina." The Journal of Neuroscience 22.7 (2002); and Pillow, Paninski, Uzzell, Simoncelli & Chichilnisky, J. Neurosci (2005).
-
The coupled generalized linear model (GLM): see Pillow, Shlens, Paninski, Sher, Litke, Chichilnisky & Simoncelli, "Spatio-temporal correlations and visual signaling in a complete neuronal population." Nature 454.7207 (2008).
All of the computational models produce a linear response, and the LNP and GLM generate spikes over a number of trials as output. The outputs can be used for a computational observer to discriminate between stimuli, and in the near future will feed into models of the LGN and V1.