How are circuits wired up during development to perform specific computations? We address this question in the retina, which comprises multiple circuits that encode different features of the visual scene, culminating in roughly 20 different types of retinal ganglion cells. Direction-selective ganglion cells (DSGCs) respond strongly to an image moving in the preferred direction and weakly to an image moving in the opposite, or null, direction. Direction-selective ganglion cells are critical for driving ocular-motor reflexes that stabilize images on the retina as we move through a visual scene as well as for sensing the movement of objects within the visual scene.

In adult retina, the preferred directions of DSGCs are not randomly distributed but cluster along distinct directions (up, down, left and right), which we refer to as cardinal axes. The mechanisms that guide the emergence of these cardinal directions and the precise excitatory and inhibitory connectivity that define them are unknown. Work from our lab and others has demonstrated that direction selective responses are detectable at the age of the earliest visual responses, indicating that the retinal circuitry mediating direction selectivity emerges prior to normal visual experience, and these responses remain after a variety of genetic and pharmacological blockades of synaptic signaling in the early retina (reviewed in 1,2). Hence the primary hypothesis has been that direction selective circuits emerge independent of neural activity.

I will present recent results from my laboratory that counters this hypothesis and demonstrates that both activity- dependent and independent mechanisms underlie the development of retinal direction selectivity.