Multiplexed neuromodulatory-type-annotated whole-brain EM-reconstruction of larval zebrafish

Electron microscopy (EM)-based reconstruction of microscopic connectomes is crucial in elucidating the synaptic organization of the brain architecture. To understand the reconstructed connectome and integrate it with functional interrogation, it is indispensable to identify neuronal types in EM datasets. Among the vast array of neuronal types, diverse evolutionarily-conserved neuromodulatory (NM) neurons endow the hardwired brain architecture with flexible neural processing and function. Here, we present Fish-X, a synapse-level, multiplexed NM-type-annotated reconstruction spanning the retina, brain and anterior spinal cord (SC) of a 6-day-old larval zebrafish, capturing >176,000 cells and >25 million synapses in the brain and SC. Among diverse neuronal types, NM (including noradrenergic (NE), dopaminergic (DA), serotonergic (5HT), and hypocretinergic) and glycinergic neurons are identified via subcellular localization of the enhanced ascorbate peroxidase-2, while excitatory (E) glutamatergic and inhibitory (I) GABAergic neurons are inferred through morphology comparison with the Zebrafish Mesoscopic Atlas. Different types of NM neurons display various synapse density and number on dendrites, with the number being positively correlated with their total projection length. By reconstructing presynaptic neurons (PNs) and identifying their E/I identity, in-depth analysis of individual locus coeruleus (LC)-NE neurons revealed that they receive brain-wide synaptic inputs, which are organized into modality- and E/I-dependent spatial patterns along dendrites. Moreover, a subset of PNs innervate multiple LC-NE neurons, and this input sharing extends across NE, DA, and 5HT systems, suggesting a co-innervation mechanism for coordinated modulation within and across diverse NM systems. Thus, our study reveals the multi-dimensional synaptic input organization of the LC-NE system. Integrated with the zebrafish mesoscopic atlases, the Fish-X offers a valuable resource for deciphering how, at the level of synaptic architecture, NM systems interact with sensorimotor pathways to shape neurocomputing in the larval zebrafish brain.