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

PCA traces have negative values #129

Open
NataliaFavila opened this issue May 3, 2023 · 3 comments
Open

PCA traces have negative values #129

NataliaFavila opened this issue May 3, 2023 · 3 comments

Comments

@NataliaFavila
Copy link

Hi Biafra,

We are wondering what exactly are we getting in the traces from the PCA-ICA analysis?
Looking at the traces stored in the PCA-ICA output.mat file we noticed some have negative values. So we are wondering whether these are raw traces or dff?

Thank you!
Natalia

@bahanonu
Copy link
Owner

bahanonu commented Jun 7, 2023

@NataliaFavila Could you either provide the PCA-ICA mat-files or detail the parameters you used when running PCA-ICA?
image

The values for PCA-ICA depend in part on the movies and whether the values were transformed to estimated movie units (e.g. dF/F) or as S.D. (e.g. see #107 (comment) or below from Fig. 2c-d from original Mukamel paper).

image

Depending on the activity of cells in the movie, there can be true negative values (e.g. inhibition after a stimulus) or negative values around zero due to noise (experimental or biological) fluctuations after dF/F calculation (e.g. if you calculated the ROI-based traces, would also see negative values). If you send a snippet of the raw movie, I can check to see which of these it likely is.

@NataliaFavila
Copy link
Author

thank you so much for the explanation, I have now run the ROI based traces which were non-negative (as we were expecting), so now it all makes sense. Thank you!

@bahanonu
Copy link
Owner

Sounds good! As a quick note, be careful about using the ROI-based traces for downstream analysis, as depending on the imaging modality, density of cells, background activity, etc. they can contain significant cross-talk compared to traces from other methods (e.g. PCA-ICA, EXTRACT, CNMF(e), etc.).

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