[ExI] HM

[Anders Sandberg]

Eugene:
> How much would we need for a mouse right now, only a few years? 
> There are a lot of cm^3 in a human brain, and scanners are expensive.

If the brain has volume V and you slice it in thickness t slices, you will get ~V^(1/3)/t slices, with a total area of V/t. 

In the case of doing a 50 nm slices of a 450 mm^3 mouse brain, this means 153,000 slices with a total area of 9 m^2. The same case with a 1400 cm^3 human brain gives 2 million slices and total area 28,000 m^2. 

Todd and the KESM team have a 2 Tb datasets of a mouse brain, but that is of course just optical resolution. Enough to see the neurons, but not enough to get connectivity.


>From a WBE perspective, the big problem is getting the scanning area up and the scanning time down. If we have N microscopes scanning an area A in time T, the total time will be V/tNAT - plus overheads due to tissue handling, presumably scaling as V^a t^-b N^c for some positive constants a, b and c (more stuff to handle, more slices to move, more destinations for slices). If A and/or T can be increased, then the total time is reduced without extra overhead. 


Anders Sandberg,
Future of Humanity Institute
Philosophy Faculty of Oxford University