Mapping of spectrally-tuned neuronal responses to visual stimuli in zebrafish – two-photon imaging and optical perturbation

Most animal species strongly depend on their sense of vision. The ability to perceive and discriminate chromatic signals is very useful for object detection and identification in the natural environment. Animals with a larval stage are generally characterized by a huge investment of energy in the rapid development of a complex retina associated with downstream visual circuits, which will improve the larva’s chances of survival with sensorial cues fundamental for both predator escape and prey capture. Indeed, the environmental distribution of radiance, covering a wide range of wavelengths, may be used to construct a visual map guiding several behaviors, based on different world representations mapped with the different wavelengths. Identifying the complex networks in charge of processing visual information in the brain and deciphering their computational activity in response to specific visual stimuli, thus, represents a fundamental task in understanding sensorimotor transduction, which is at the basis of many behaviours. The transparent larva of zebrafish represents an ideal model for imaging neurons and their activity throughout the whole central nervous system (CNS) of the animal, enabling the study of the neuronal circuit correlates of behaviors. The visually-guided behavioral repertoire of larvae (e.g. optomotor and optokinetic responses, phototaxis and prey capture) becomes quite rich just after few days of development. In this context, color vision, based on a spatially anisotropic tetrachromatic retina, provides a crucial evolutionary aspect of world representation, driving some fundamental larval behaviors. Color information is integrated and processed at various levels in the retina (where specialized cells are involved in encoding spectral signals deriving from different part of the visual landscape) before conveying this information to the brain, which is the object of study of this thesis. To investigate the processing of chromatic information in the zebrafish larval brain, we mapped, with cellular resolution, spectrally-responsive neurons in the larva encephalon and spinal cord. We employed the genetically-encoded calcium indicator H2B-GCaMP6s and two-photon microscopy to image neuronal activity in zebrafish larvae at two different stages of development (3 and 5 days post fertilization, dpf) while performing visual stimulation with spectrally-distinct stimuli at wavelengths matching the absorption peaks of the four zebrafish retina cone types. We implemented segmentation and regression analysis to identify neurons selectively responding to one or multiple stimuli and we registered them onto a reference brain for precise anatomical localization. These measurements revealed the presence of a high number of wavelength-selective neurons not only in the brain areas directly involved in processing visual information (for example the optic tectum and the other retinorecipient areas), but also in all other regions of the CNS down to the spinal cord. The zebrafish larva is a very simple model yet representative of vertebrate CNS organization, so we expect the paradigm of spectral information propagating through non-visual areas to be relevant in all vertebrates. Comparison across larvae at two different stages of development highlighted different spectral patterns of activity with responses strongly dominated by UV stimuli in 5 dpf larvae and red stimuli in 3 dpf larvae. The latter showed a negligible level of integration between multiple stimuli and no wavelength-selective neurons in the spinal cord. The spectral tuning of responses and anatomical maps of neurons involved in color-driven behaviors indicate the complexity of the circuits involved and open the way to their detailed investigation. Perturbative optical approaches, based on laser ablation and optogenetics techniques, allow dissecting neuronal circuits underlying visually-evoked responses. As part of this work, I setup novel methods for GCaMP-based temporary and selective perturbation of neurons and for optical highlighting of their intricate neurite structure. These methods were tested with the inactivation of tectal neurons responding to specific spectral stimuli, followed by a mapping of their projections into the tectal neuropil. This approach, in combination with the pan-neuronal expression of the genetically-encoded Ca2+ indicator GCaMP6s, allowed us to induce a temporary loss of function to the visual stimuli specifically localized to the irradiated cell, without affecting the surrounding neurons. Therefore, the method developed in this thesis can be applied to the study of local connectivity and networks in the CNS. Understanding the workings of the vertebrate brain is one of the main challenges of modern science. In this context, the development of novel approaches at the interface between physics and biology drives constant improvements in our capability of studying neuronal activity at large scale and dissecting connectivity patterns to form complex and dynamic networks. The goal of measuring neuronal activity in real time and in an entire vertebrate brain represents a formidable task and finds in the transparent zebrafish larva the best model for implementing many innovative technologies. In this thesis, two-photon microscopy allowed to investigate the distribution of spectral information across the whole brain and spinal cord, providing novel insights in this field and laying the ground for future investigations based on perturbative approaches.