Raw fluorescence vs. dF/F

[PTRRupprecht]

First, thanks a lot for making your datasets publicly available!

I had a brief look into the calcium imaging data set, and I have a very short and specific question: The data set includes the "raw fluorescence", as you write in the Readme:

nda.Fluorescence: Records the raw fluorescence traces for each segmented mask

However, there are also negative values, and the range of the traces seems to be closer to dF/F, rather than raw fluorescence values from the microscope. Could you clarify whether the nda.Fluorescence is raw fluorescence or "raw" dF/F ?

Best wishes,
Peter

[jakereimer]

Hi Peter, sorry I didn't see this earlier. The fluorescence traces from the segmented masks are fluorescence traces output from CaImAn. The acquisition system (ScanImage) collects values as signed int16 and there is some arbitrary offset that can result in negative values.

[PTRRupprecht]

Hi Jake, thanks for your reply!
That is a bit unexpected. In the current format, the fluorescence traces have a mean which is very close to zero (compared to the standard deviation of the same signal).
If there is indeed an arbitrary (and uncompensated) offset to the signal due to Scanimage, this would prevent the proper computation of dF/F values.
Another possibility which comes to my mind is that CaImAn might subtract the background components, thereby reducing the baseline F0 of the extracted components to almost zero.
Both cases would not be ideal if you want to perform analyses based on the absolute values of dF/F. But if you are not interested in that, it would of course not make a difference to you.

[jakereimer]

Hi Peter, that's right CaImAn does subtract a background component.