Showcase Symposium 2020 Project Talk Q&A: Visual Coding Neuropixels Pipeline

Show & Tell
,
1

At our virtual Showcase Symposium 2020, Allen Institute for Brain Science teams present project talks highlighting their work. Please use this forum thread to ask questions for the speakers of the Visual Coding Neuropixels talk with @joshs, Severine Durand, Tamina Ramirez, @rami, @jxiaoxuan, Zack Kilpatrick (University of Colorado), Corbett Bennett, and Gregg Heller. Please see the full list of Showcase 2020 forum threads to ask questions about other talks.

2

Question from the Showcase Symposium Q&A from Anne Churchland: Super interesting stuff! I had a question about the manual adjusment of boundaries after alignement to the CCF. Are there sometimes outliers? That is, does a probe sometimes not end up where it is supposed to? How are these identified and are there agreed-upon criteria for excluding such recordings?

3

@rami, can you explain more what the subspace angle is measuring?

4

This question is for Rami Iyer.
Did you consider a statistical null, e.g. permutation test to compare with expected subspace angle under randomized conditions?

5

Question for Ram from @farznaj, who is having technical difficulties: I didn’t quite get how you concluded from the angle analysis that lower dimensions are carrying global information, while the higher dimensions are specific information. Could you please elaborate?

6

Question for Ram: I didn’t quite get how you concluded from the angle analysis that lower dimensions are carrying global information, while the higher dimensions ar specific information. Could you please elaborate?

7

Thanks for the question, Anne! Because the insertions angles are fixed on our rigs, the probes travel along highly stereotyped trajectories, and there aren’t really any outliers. For each of the 6 probes in a given experiment, we’ve become very familiar with the relevant physiological landmarks, which makes the manual adjustment straightforward. That said, we currently rely on “major” boundaries (cortex/hippocampus, hippocampus/thalamus, or hippocampus/midbrain), and do not have metrics for refining boundaries between individual cortical layers, hippocampal subfields, or thalamic nuclei.