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

Sorting parameters #74

Open
bgu1997 opened this issue Apr 22, 2023 · 1 comment
Open

Sorting parameters #74

bgu1997 opened this issue Apr 22, 2023 · 1 comment

Comments

@bgu1997
Copy link

bgu1997 commented Apr 22, 2023

Hi,

  1. I noticed there was a discrepancy between the Wiki and what is in the options.py file with regards to the sorting parameters. On the Wiki, there is a parameter named "MaxDistGrouping," but in options.py, this parameter is (presumably) named "MaxDistMatchGrouping." Can you please confirm that these are the same parameters?
  2. As suggested in the Wiki, I found that Combinato with default parameters overclustered on longer recordings (~10 hours) where it often found 20-100 units per channel. After some experimentation, increasing MaxDistMatchGrouping from 1.8 to 2.5 and increasing MinSpikesPerClusterMultiSelect from 15 to 25 seemed to be a good middle ground. However, I am still getting >15 units on many channels with many more valid clusters starting out in the Artifacts bin. What is the reasonable range of parameter values for Combinato?
    Thank you!
@jniediek
Copy link
Owner

Hi!

  1. Regarding the parameter name, you are absolutely right. The correct name is MaxDistMatchGrouping. The parameter is used only at one code point in all of combinato, here:

crit = options['MaxDistMatchGrouping']

I updated the Wiki, thank you very much for pointing this out!

  1. a) Great to see that you already experimented with parameters. What you describe is the normal variability in 10 hour recordings. In my experience, it is worthwhile using the GUI to merge clusters in these cases. Are you aware of the `Ctrl+1 shortcut? It merges all groups that consist of exactly one cluster, which often makes the manual sorting more efficient.

    b) If you see many true clusters in the artifacts, you might want to play with the artifact cluster parameters. They are defined in options.py:

    artifact_criteria = {

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

No branches or pull requests

2 participants