About 5 years ago, neuroscientist Tony Zador was struck by a novel idea for how to build a synapse-by-synapse map of brain connections—a goal dear to neuroscientists who want to understand how the brain's intricate wiring underlies its functions. By using random series of 30 nucleotides—the building blocks of RNA and DNA—he thought he could give neurons unique barcodes, linking the barcodes across synapses, then create maps of their connections using high-throughput gene sequencing technology. Now, after a long "slog" of technical difficulties, the researcher at Cold Spring Harbor Laboratory in New York believes the technique is ready for prime time. As part of a project led by molecular engineer George Church of Harvard University, the technique was selected last month to receive a $21 million, 5-year brain mapping grant, one of six new projects under the Machine Intelligence from Cortical Networks, or MICrONS, program, sponsored by the Intelligence Advanced Research Projects Activity. Ultimately, MICrONS aims to map every neuron and synaptic connection within a 1-cubic-millimeter chunk of tissue from the mouse visual cortex, and use those detailed brain connection maps to design computer architectures able to perform tasks that are easy for a brain but out of reach for artificial intelligence.

Craziest idea I've heard in 2016, but who knows, it might have a chance of working. FISSEQ alone took a decade to get working, but in principle, this pitch is sound...