Skip to main content

In the eye of the observer: Visual processing at the heart of the retina

Periodic Reporting for period 2 - switchBoard (In the eye of the observer: Visual processing at the heart of the retina)

Reporting period: 2017-11-01 to 2019-10-31

A key organisational principle of the visual system is parallel information processing, which starts to emerge already in the retina, the thin nervous tissue that lines the back of the eye. Ontologically a part of the brain, the retina is the first outpost of visual processing. In contrast to other senses, vision relies on a sensory organ equipped with its own neural processor. Instead of “wiring” the sensory cells directly to neurons that project to the brain, the retina contains an additional layer of interneurons, which heavily process the light signal from the photoreceptors before it is passed onto the retinal ganglion cells (RGCs) – the retina’s output neurons – and forwarded via the optic nerve to higher visual brain centres.

At first glance, the organisation of the retina looks simple: An excitatory signal pathway links the photoreceptors via the bipolar cells (BCs) to the RGCs. Signals along this “vertical” pathway are shaped by lateral inhibition provided by horizontal cells (HCs) and amacrine cells (ACs) in the outer and inner retina, respectively. The surprising diversity of neuron types, however, hints at the complexity of retinal signal processing: The signals from 3-4 photoreceptor types – modulated by 2-3 HC types – is distributed onto more than a dozen BC types, indicating that parallelisation of visual information already starts at the first synapse of the visual system. In the inner retina, the BC signals are picked up by the dendrites of around 45 types of RGC. Moreover, likely a similar number of AC types shape the signal flow from BCs to RGCs via highly selective synaptic interactions. Finally, each RGC type relays a separate representation (“view”) of the visual scene to the brain.
At the heart of retinal signal processing lies a thick and dense synaptic layer, where the axon terminals of BCs interact with the dendrites of ACs and RGCs. With its layered organisation and highly selective connectivity, this so-called inner plexiform layer is reminiscent of an old-style electric switchboard for managing phone lines – hence the name of the consortium. Because of their central position in the retinal network, BCs represent a unique starting point for unravelling key principles of parallel processing: First, they implement the first stage of signal parallelisation in the visual system. Second, BCs provide the excitation that drives the extraction of visual features by the inner retinal circuits, thereby forming the basis for the next, much larger set of parallel information channels represented by the RGCs. Third, with “only” around 15 types, BCs are sufficiently diverse while experimentally well approachable.

The overall scientific objective of switchBoard aimed at a comprehensive understanding of BCs, their functional organisation, and their role in the first critical steps of vision in health and disease. The consortium’s combined expertise in neuroscience and vision research, together with the exceptionally broad spectrum of cutting-edge methods in the partners’ labs, enabled the early stage researchers (ESRs) to bring us through their work an important step closer to this goal.

switchBoard’s scientific objectives went hand in hand with its main training goal: To prepare ESRs for a successful career in a quickly changing research field. Neurosciences offer attractive interdisciplinarity to ESRs, with possible career paths in both the public and private sector. For a successful career in neuroscience, however, ESRs must be trained in multiple fields. Consequently, switchBoard ensured that ESRs received in-depth training in experimental and computational neuroscience, neurotechnology, and biomedicine. To this end, the consortium implemented an intense training programme, complemented by hand-on workshops organised by all our private sector partners. Through its interwoven research and training program, switchBoard contributed to replenishing resources that are often taken for granted but are of paramount importance for Europe: by training the next generation of competitive, multidisciplinary young scientists and by generating knowledge through basic research.
Work package (WP) 1 investigated different aspects of BC function in the healthy retina. The ESRs revealed, for instance, how morphology, channel complement and possibly electrical coupling to ACs may jointly shape the cell’s synaptic output, that BCs use multi-vesicular synaptic release to dynamically adapt the way they encode information, and that the diverse colour responses in zebrafish BCs are evolutionary matched to the animal’s environment. WP2 investigated how retinal diseases, like photoreceptor degeneration and synaptic protein deficiencies, alter BC function and, consecutively, affect vision. Here, an important result was that photoreceptor degeneration may trigger remodelling of inner retinal circuits to a previously unexpected degree. In addition, ESRs found that BCs play a central role in mediating aberrant retinal activity that may even be linked to pathological eye movements in human patients. WP3 developed computational models of BC function and new experimental technologies. These efforts resulted, for example, in a systems identification model that shows a much higher predictive performance than “traditional” methods, and in biophysically realistic models of single BCs that simulate, for instance, the cell’s responses to electric fields, mimicking the effects of stimulation by a subretinal implant. In addition, ESRs identified new applications for retina-inspired cameras and made substantial progress in the development of electrical and optical imaging approaches using high-density multi-electrode arrays (MEAs) and digital mirror devices (DMDs), respectively.
The modelling efforts in WP1 and WP3 yielded for most BC types a range of biophysically realistic models. Some of these also simulate transmitter release, which will allow researchers to compare experimental with modelling data at the level of the cells’ output to the inner retina and helps integrating BCs into larger retina models. In addition, WP3 developed “batch” correction methods to facilitate across-experiment data integrating. We expect that these resources and the ESRs trained at the interface between computational and experimental neuroscience will help bringing these field closer together. The results of WP2 strengthened the view that BCs are key to neuroprotective interventions as well as vision rescue approaches. At the same time, the ESRs found that BC are critically involved in modulating aberrant activity that limits the effectiveness of such treatments. Moreover, WP2 demonstrated how to track the effects of a BC-specific mutation from the circuit to the behavioural level in the mouse model, and to tie it to the phenotype in human patients. Hence, we expect that WP2’s results significantly contribute to the knowledge that is instrumental for devising future treatments.

Aside from the research, the biggest asset generated by switchBoard are certainly the ESRs, who thanks to the broad expertise provided by academic and private sector partners, received an intense multidisciplinary neuroscience training, which enables them to follow career paths in academia and industry.