Tuesday, November 9, 2021

Notes

Online 3D mouse brain viewer here https://connectivity.brain-map.org/3d-viewer
Should consider putting some of these in thesis to help explain tractography etc.

I have realised that I messed up the explanation as to why Monte Carlo network is useful.
The Monte Carlo network can EXACTLY model gradients, layers, etc. etc.
As such the stats model is designed to make this faster, since the Monte Carlo can already do it - it just takes forever.
With this in mind, I should design a version of the experiment where layers and gradients etc. etc. are actually involved in the problem.
Then check, with these, how close the summary version (the stats) matches to the actual problem.

What did the first reviewer mean by not knowing the out-degree, I thought I already handled this (the outdegree follows a random distribution).
Why more interesting to not know the out-degree as opposed to the in degree also??