Vision and the evolution of neuronal computation

We use the vertebrate retina as a central discovery platform for understanding the evolution of computation in the nervous system.

How do synapses, neurons and networks enable ‘function’, and how can they rearrange to meet new sensory and behavioural demands on evolutionary timescales? To address these questions, we link state-of-the-art approaches in systems and computational neuroscience with behavioural and sensory ecology, while also drawing on connectomics, transcriptomics and genetic manipulation. Our model systems include zebrafish and mice, as well as birds, frogs, and sharks. Our general trajectory is elaborated in recent reviews (Nature Rev Neurosci 2020, Curr Biol 2021) and a special issue on “Vision in non-model vertebrates” (SDCB 2020).

Evolution of vision

The retinas of all vertebrates are based on a common blueprint but different species evolved a vast range of retinal particularities that help them  navigate their visuo-ecological niche. How are these differences reflected in retinal networks? The answer towards this, and many other questions relating to animal lifestyle and evolutionary history must come from comparisons across species, including non-model species. We therefore complement our core work on zebrafish vision with frequent excursions into the sensory complement of other animals.

Vertebrate colour vision

With the rise of the dinosaurs, our own ancient ancestral proto-mammals escaped to the forests and adopted a nocturnal lifestyle. In this process, they lost all but two of their cone-photoreceptor types, the basis of colour vision. As a result, most of today’s mammals are dichromats, or “atypical trichromats” such as our own species. Not so the zebrafish, or indeed almost any other vertebrate: Most birds, fish, reptiles, amphibians and even jawless fish still “see in 4 or even 5 colours”, based on the 4-5 cone types that were around long before the dinosaurs took over the planet.

Synaptic computation

The computational power of retinal networks is deeply rooted in their synaptic architecture. The highly specialised ribbon-type synapses of photoreceptors and bipolar cells receive a myriad of feedback and feed-forward signals through the network, making them probably the most important computational locus of the circuit. Using 2-photon imaging of voltage, calcium and synaptic release signals in vivo we aim to probe the visual response properties of retinal synapses towards a better understanding of locally visual signal transformations. 

The natural visual world of animals

All sensory systems are specialised to best serve an animal’s sensory-ecological niche, and vision is no exception. Here, specialisations range from basic anatomy all the way down to the tuning individual synapses and microcircuits. To understand what natural sensory demands and pressures shape requires measurements of natural visual worlds, including systematic variations in space, time and wavelength composition. We use custom camera and hyperspectral scan-systems to survey visual information available to animals in nature.

Retinal coding

The retina breaks light patterns into increasingly specific internal representations of the visual world. Spike trains of retinal gamglion cells (RGCs), the retina’s only output neurons, retain only a tiny fraction of original visual input to form a sparse but information rich retinal code. Within the retina, interactions between five principal neuron classes, each comprised of several types, fundamentally shape the response properties of individual pathway elements. We study the gradual evolution of feature extraction as visual information trickles through the retinal network.

Technology development

To probe neuronal function in the living nervous system requires state of the art imaging equipment such as our heavily customised 2-photon microscopes and visual stimulators. Here, we aim to drive current possibilities in optical imaging technology, for example by light-path optimisations or the implementation of custom scan strategies that acknowledge the 3D curvature of biological samples. In parallel, we invest heavily in open hardware approaches to optimise research-grade lab equipment for a fraction of commerical costs.

Why retina?

The retina is probably the most accessible part of the central nervous system. Its entire computational logic can be interrogated in a dish, from patterns of lights as the natural input, to spike trains on the optic nerve as the natural output. Consequently, retinal circuits include some of the best understood computational networks in neuroscience.

The retina is also ancient, and central to the emergence of neurally complex life on our planet. Alongside new locomotor strategies, the parallel evolution of image forming vision in vertebrate and invertebrate lineages is thought to have driven speciation during the Cambrian. This early investment in sophisticated vision is evident in the fossil record and from comparing the retina’s structural make up in extant species. Animals as diverse as eagles and lampreys share the same retinal make up of five classes of neurons, arranged into three nuclear layers flanking two synaptic layers. In fact, some retinal neuron types can be linked across the entire vertebrate tree of life. For example, the photoreceptors in our own eyes can be directly linked to those in eagles and lampreys. And yet, the functions that these photoreceptors serve in each species, and the circuits that they innervate to do so, are almost certainly distinct to acknowledge the vast differences in species-specific visuo-behavioural demands.

 

The future

Despite a nuanced understanding of mostly mammalian retinas, and an emerging point of comparison in zebrafish, we still know essentially nothing about circuit functions of most non-mammalian eyes. And yet, many of these animals far surpass the visual abilities of mammals by any measure of the term. For example, species of diurnal raptors such as eagles or falcons have the highest spatial and temporal resolution of any vertebrate. This allows them to stabilise flight at velocities that would put a cheetah to shame, or to spot near imperceptible detail on the ground from 100s of meters up in the air. They combine these abilities with a sense of colour vision that long surpasses our own.

However, we only have to go as far as our own backyard to encounter key circuit-level differences to those of mammalian retinas. For example, even the retina of the flightless domesticated chicken is nearly twice as thick as that of mice, comprising an order of magnitude more neurons per equivalent area. What do all these ‘extra’ neurons do, and how are their circuits linked to those of mammals? Here, where birds provide a key point of comparison regarding complexity, other lineages offer the possibility to understand how and when in evolutionary time diverse circuits emerged. For example, of the well-studied motifs in mice, which are also present in sharks, the earliest extant lineage of jawed vertebrates? Or how can circuits for colour vision – probably the oldest of all retinal circuit motifs – be linked across species, perhaps predating the emergence of vertebrates themselves?

To address these types of broad and open-ended questions requires a research programme willing to take substantial long-term risks. However, the pay-off may be equally substantial. Technology enabling large-scale recordings from neuronal populations are rapidly coming in reach for species beyond the traditional few that grant straightforward genetic access, while sweeping developments in molecular and anatomical techniques are starting to enable truly comprehensive surveys on the make-up of neuronal circuits. Therefore, now is the time to seriously invest in a broad approach – both in terms of species, and in technology – that will in time allow us to bring modern systems neuroscience into the wild. And what better model system to use when allowing these possibilities to flourish than the vertebrate retina?