Tony Zador of Cold Spring Harbor Laboratory devised a new technique for mapping connections among neurons. It is much faster than other methods and at least as accurate as the most accurate competing methods, including fluorescence techniques. The technique, MAPseq, uses genetically modified viruses to insert unique RNA sequences (“bar codes”) into each neuron. Post-mortem DNA sequencing identifies connections among all neurons in the sample. The resulting model is structural, not functional. Derived models are not spatially accurate (i.e., not to scale and not physiographically representative). The models identify intraneural connections but not specific messaging among neurons. Zador is pursuing functional analysis by combining MAPseq with other techniques. MAPseq currently can map about 100,000 neurons per week. Increasing hardware and software efficiency and power will improve throughput dramatically over time.

This is the most startling brain research development Mark has come across recently. The implications are tantalizing. Start with embedding unique codes (think of inventory numbers) in each neuron. Presumably using a virus to add a consistent unique identifier to every cell in an organism could result in a unique  “bar code” for every human and every other organism. We already have such a code in our genome, but this method could create a simpler code that would be easily readable by miniature, portable DNA sequencers. It could be a shorthand code linked to a person’s full genome record. 

Back to brain research, once Zador and others find ways to combine real-time functional mapping and non-destructive ‘reading’ of the cellular IDs, increasingly faster computing and smarter (AI-enabled) software may make it possible to map not only a person’s neural connectome, but the functional dynamics playing out in the brain from moment to moment. That, in turn, could make it possible to create a high-fidelity, functional copy of a human mind (aka, a ‘mindclone’). It would probably not be necessary to explicitly model every neuron, synapse, and intraneural communication, but that may one day be possible.