Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

ChanMap_24LinearUprobe #155

Closed
Stefxnia83 opened this issue Dec 6, 2019 · 7 comments · Fixed by #595
Closed

ChanMap_24LinearUprobe #155

Stefxnia83 opened this issue Dec 6, 2019 · 7 comments · Fixed by #595

Comments

@Stefxnia83
Copy link

Hi,
I created the channel map for a linear Uprobe (24 channels spaced 100 micron) to sort Plexon data with Kilosort 2. I set the kcoord=1 for all the channels [1,1,1..1], as i have just one shank.
My question is: is that correct? or I should put kcoord=1:24 [1,2,3,..24]?
The confusion comes from the fact is not clear to me, if the algorithms uses this information to:

  1. decide if two clusters labeled as good and belonging two different channels close one to each other, are actually generated by the same source, so kilosort return just once;
    or 2) whiten and clean the signal.

Thanks,
Stefania

@marius10p
Copy link
Contributor

Kilosort and Kilosort2 by default sort all channels together, so there is no problem of matching units sorted from different channels. "kcoords" is currently inactive in Kilosort2, and was used in Kilosort1 to indicate sets of channels with completely non-overlapping waveforms. In all cases, nearby channels are used to whiten the signal, even if they are very far apart.

@Stefxnia83
Copy link
Author

Thanks for your reply, it's very helpful!

@JoseGuzman
Copy link
Contributor

Is there still any interest to implement kcoods in Kilosort2? Originally posted by @marius10p in #21 (comment)

Or is this a feature that is not going to be implemented?

@marius10p
Copy link
Contributor

I think there is a lot of interest, I just don't have any time. Any takers? The code is now commented at a fine level of granularity.

@JoseGuzman
Copy link
Contributor

Would you give us a hint of where to look (file, lines) ? I'm new to matlab, and may need to become familiar with it first.

@marius10p
Copy link
Contributor

I think a good place to put it might be in triageTemplates2, which gets called here. and here. You would have to mask dWU, which is the running average of the waveforms (aka the "template"), as well as "U", which is the spatial component of the low-rank decomposition of dWU. Probably no need to change anything else, but I am not 100% sure... If you are more familiar with Python, you could try to start from this version, which is a very close, function by function port that is now nearing release.

@marius10p
Copy link
Contributor

Closing for lack of activity.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging a pull request may close this issue.

3 participants