Skip to content

Latest commit

 

History

History
85 lines (65 loc) · 2.99 KB

notes.md

File metadata and controls

85 lines (65 loc) · 2.99 KB
Raw

questions

  • input / output / inner
    • what
  • keep only A->D neurons
    • neurons or edges?
  • identify "latent structure" (clustering, should be feedforward)
    • how shoulw we constrain the clustering to be "feedforward"?
  • then add others synapse types, how does structure change?

  • look at PN -> KC "pattern"

  • is that "pattern" present elsewhere?

  • where in the brain do L and R NOT match

  • central complex

    • why do they care about this?

  • circuit motif discovery

    • how?

  • are there any other groups of neurons "like" kenyon cells

  • eg, that they can't be uniquely identified for some reason (eg, doesn't exist on other hemisphere)

    • most other cells are individually matched across brain.
  • we can directly send results also to AC/MZ
    • I think I have been?
  • ch 1: 3 layers, what structure is in there?
    • what are the 3 layers?
  • ch 2: 4 colors, what kind of circuit motifs are there together and different?

  • how is larval dataset different from adult dataset??

    • does this mean we will have the adult data?

  • cross hemispheric differences

  • inter-hemisphere

  • synapse type differences

  • multi-sensory/inner/motor (the adults don't have the complete proprioceptive)

  • efference copy / proprioceptive

do we have the labels? modality? input? output? - we know sensory modality for PN and thats all they have given us

help AC a lot in short term:

  • color the pathways, these are 21 ORNs, etc. color each group,

    • what does color each group mean

  • see the merged colors, which neurons are getting input from each of these types of inputs,

    • I could certainly look at the proprotion of input onto a non PN from each PN type

  • until eventually reaching the command neurons?

  • how is MB injected into these paths of convergence?

  • AC expects: 2 groups, regular sensory, the other is mostly proprioceptive which has its own circuits

    • proprioceptive should be coming from the other direction?

  • the "larvunculous".

    • what's that

  • closer to output layer proprioceptive convergences with other sensory

  • there are 12 segments, somatosensory neurons is largest, each segment with 40 per hemisphere, so 40x2x12 = 100 neurons!

    • are you talking about lower in the body?

  • efference copies:

    • the descending neurons which also project into the brain?
    • neurons that synapse onto the descending neurons?
    • eg, output and input of descending neurons

  • stratified thing, AC expects layers (only clear layers are input and output layers), between these layers there will be feedback. much of this feedback is A->A, that is, subsequent layer impacts previous layer via A->A

    • why do they think this?

  • how many layers? maybe 6 or 7, but there are shortcuts.

  • what is the diameter of the graph? expectation: 10 for A->D.

  • feedback?

  • are there hubs? is it a powerlaw?w

    • powerlaw degree distribution?

  • we can also compare to c. elegans!

    • cool, scott's?

  • Q1: what is info flow from input to output?

  • Q2: you can have many parallel pathways per layer?