search

kushalkolar / MESmerize

Platform for Calcium Imaging analysis. DEPRECATED.
GNU General Public License v3.0
58 stars 14 forks source link

Improve CNMF with evenly lighted recordings? #75

Closed TundeSzungyi closed 2 years ago

TundeSzungyi commented 2 years ago

Many false negative and false positive results of CNMF

Hi,

I have a problem, that even though I tried very different parameters with CMNF some neurons are not found and at some places, there is a very big amount marked, which creates false-positive results. I attached my photo of my CNMF results (I tried different neuron sizes, SNR, neuron numbers per patch).

The boundaries of the miniscope recordings are darker and I was thinking that maybe this is the problem, because I got stuck that so many neurons are not found.

Do I need to make some preprocessing of the videos, are there some methods that I could use? Or should I look for parts that have the same lightning and crop those parts to analyze further?

Here is my screenshot: image

Thank you very much in advance

kushalkolar commented 2 years ago

You will need to use CNMFE for 1p data.

TundeSzungyi commented 2 years ago

Thank you!