Abstract
Purpose:
In mammals, retinal ganglion cells (RGCs) are a heterogeneous population, with each type thought to encode a different aspect of the visual scene. The mouse retina is an attractive model for studying retinal function and development because it can be readily manipulated on a molecular level. However, comprehensive functional characterization of mouse visual pathways is needed to take the full advantage of these manipulations. Here we describe our progress in classifying mouse RGCs through their responses to various visual stimuli.
Methods:
RGC responses were measured by placing an isolated wildtype mouse retina on a 512-electrode array and projecting various visual stimuli, such as spatio-temporal white noise and drifting gratings, onto the photoreceptor layer. Action potentials from hundreds of RGCs were identified in a single retinal preparation. Classification of RGCs within a single retina was based on their functional properties such as spike-triggered average response to white noise stimulus. Principal Component Analysis was used to extract features of the response most useful for classification. Spatial organization of the receptive fields was not taken into account in the classification process. RGC types with similar response properties were found across multiple retinas by clustering the RGC types identified in the individual preparations, while attempting to correct for retina-to-retina variability.
Results:
Our data shows that a number of individual RGC types can be consistently identified in wildtype mouse retinas and that this identification can be done exclusively through their response properties to visual stimuli. These cell types include ON- and OFF-center RGC types with distinct spatio-temporal filtering properties and direction selective RGC types. Furthermore, we find that individual cell types’ receptive fields tile the retina in a regular mosaic pattern.
Conclusions:
We show that RGC types in the mouse have distinct functional properties, which are consistent across different retinas. This indicates that each RGC type represents a distinct retinal circuit that sends information about a particular aspect of the visual scene to the brain. The presented functional characterization and classification will be valuable for investigating retinal function and development. It provides a necessary control and a toolbox for identifying changes of the complex retinal structure and function in genetically manipulated mice.
Keywords: 693 retinal connections, networks, circuitry •
691 retina: proximal (bipolar, amacrine, and ganglion cells) •
758 visual fields