A Cellular-Resolution Atlas of the Larval Zebrafish Brain

Abstract

Understanding brain-wide neuronal dynamics requires a detailed map of the underlying circuit architecture. We built an interactive cellular-resolution atlas of the zebrafish brain at six days post-fertilization (6 dpf) based on the reconstructions of over 2,000 individually GFP-labeled neurons. We clustered our dataset in 'cell types', establishing a unique database of quantitatively described neuronal morphologies together with their spatial coordinates in vivo. Over a hundred transgene expression patterns (mainly Gal4 lines) were imaged separately and co-registered with the single-neuron atlas. By annotating 73 non-overlapping brain regions, we generated from our dataset an inter-areal wiring diagram of the larval brain, which serves as ground truth for synapse-scale, electron microscopic reconstructions. Graph theoretical analysis identified statistical properties of this network, including hierarchically organized modules interlinked by central information hubs. Interrogating our atlas by 'virtual tract tracing' has already revealed previously unknown topographical relationships among sensory afferents in the tectum and between the cerebellum and its multiple projection targets. In conclusion, we present here an evolving computational resource and visualization tool, which will be essential to map function to structure in a vertebrate brain.