Mapping brains at arbitrary scales from the cellular composition of entire brains to reconstructing every synapse is a powerful tool for understanding brain function and dysfunction. However, the financial, engineering and computational barriers to new brain mapping technologies are far beyond the reach of most laboratories and even universities and institutions. A NeruoHub at a national lab, Argonne National Lab (ANL), could provide widespread access to such technologies. Specifically, the Hub will leverage the brightest hard X-ray source and one of the fastest computers at ANL, to provide mesoscale maps that detail the locations of every cell, their shapes and ultimately their connections with each other in large volumes to entire brains. The Hub will craft these technologies for brain mapping across diverse species, specifically the nervous system of the octopus. Enabling comparative studies across brain regions and species like the octopus will provide insights into what is general and what is idiosyncratic, and can reveal novel solutions when applied to species with distinctive perceptual or behavioral capabilities. Finally, the Hub will nurture and develop the relationship between neuroscience and the national lab system. For decades, many other fields of science have benefitted from strong collaborative relationships with the national lab system, leveraging large-scale resources to significantly advance their respective fields. Neuroscience has not. The Hub will establish the first pipeline for providing access and introduction to the broader neuroscience and national lab communities, paving the way for future collaborations.

The technical approach is to develop sample preparation protocols for staining large volumes to entire octopus brains with heavy metals (e.g. osmium, lead, and uranium), embedding in plastic, and imaging with synchrotron source micro- and nano-x-ray microscopy followed by automated serial section electron microscopy. Development of new protocols will leverage existing protocols and imaging approaches for mammalian brains while also developing capabilities at ANL for 100nm resolution projection x-ray microscopy. Where possible, reconstructions of neurons and their processes from micro- and nano-X-ray datasets will be validated with automated large volume serial electron microscopy. In tandem existing machine learning algorithms used for tracing neurons and their processes will be scaled onto Argonne High Performance Computers for analyses of X-ray datasets.