David A. Markowitz

Jeff Licthmann gave recently a talk where he outlined his vision for delivering the mouse synaptic-level #connectome from nanometer-scale volume electron microscopy with multi-beam SEM. It wasn't cheap, but feasible.

My EM colleagues have expressed concerns about feasibility. If 1 in 3 samples yields a usable 1 mm^3 volume (based on MICrONS experience), and you need ~400 such volumes/mouse, then Pr(all volumes are good in 1 mouse) is vanishingly small.

1. I would guess basic bottleneck of tissue processing and handling Petabytes of data. 2. Not many, IMO. Would need n=100 before we even start to have a chance of learning something univ. 3. A good way of doing pattern analysis on multi-individual connectomes. (DTI methods?)

3). Inverse plasticity. I give you a connector and the relevant statistics of the world. I want to solve for what the brain is optimizing. It may be possible.

Do you think this is possible in the absence of functional annotations?

I think so. But it is hard.

We tried. Tough to infer connections, and to infer *changes* in connections is way harder. Must measure many synapses at 2 times.

I think @SebastianSeung has been thinking about potential ways around this problem...

Would be very curious to see his ideas.

I may come across as molecular annotation extremist, but I bet with enough static molecular detail one could get something indicating dS/dt where S is synaptic strength.

supposing one has only a coarse anatomical "correlate" of strength (number of synapses; cumulative area of contact maybe) how good can one do at predicting circuit behavior(s)? iow, are there any cheeky shortcuts where we can do "good enough" without knowing precise strengths?

Hey @BWJones you want to start 2019 off with a thread? Next level challenge see if you can make it three tweets before saying the words "gap junction"

Nnnnnnnngh... Ghhhhhnnng..... GAP JUNCTIONS! Nope... couldn’t do it.

Weak-sauce, TEMaster. Weak.

I figure that if you used 2nm/pixel TEM (our operating resolution 2x2x70), a (1mm)^3 of brain would be 3500PB. 2nm/px is the nominal resolution required to see gap junctions which are critical to creating the edge node graph for understanding networks.

“A connectome is a complete graph of a neural network. In principle, it is not an approximation or even a statistical average. It is a comprehensive list of every connection in a defined neural region.” So, to answer your question, we get neural topology.