Visual information processing begins in the retina, a thin neuronal tissue lining the back of the eyeball. As a part of the brain, the retina does not only convert the incoming stream of photons into electrical signals, it also performs a detailed and highly specific analysis of the observed scene. Therefore, the retina can be considered a highly specialized and sophisticated image processor.

All visual information sent from the retina to the brain travels along the optic nerve, a major bottleneck of the visual system. Therefore, prior to transmission to the brain, important aspects of the observed scene must be extracted and encoded as spike patterns. These features include simple ones such as contrast, brightness, and “colour”, but also more complex ones, such as information about objects moving relative to the background. Thus, the retina sends in parallel many representations of the visual scene to the brain; each of these representations encodes different features and is represented by one of the roughly 40 retinal ganglion cell types whose axon form the optic nerve. The importance of retinal signal processing is highlighted by the fact that important decisions – what visual information is relevant, and what can be safely discarded – is made already in the retina.

The computational capabilities of this intricate neuronal network rely on more than 100 types of retinal neurons organized in complex microcircuits. Our work aims at unravelling function and organization of retinal microcircuits towards a better understanding of the underlying computational principles. Furthermore, we are interested in how these circuits are altered during degeneration.