My suggestion would be to try and find an imaging condition to minimize or prevent bleaching. Try cutting your illumination intensity way down and increase the gain of your camera. If you use a very sensitive EMCCD or one of the new sCMOS cameras with better quantum efficiency you can potentially use very little light for imaging. This is a far better solution than to try and correct for it.

Hi Blandine,

I think Christopher is right, but I would choose gain incresements only as the second choice. First I would try to work with pixel binning. Usually you do not need a high spacial resolution, and in contrast to gain binning will not raise the noise.

Check this:

https://www.microscopyu.com/techniques/fluorescence/optical-system-and-detector-requirements

Best Soenke

Hello,

Thank you for your answers. Indeed, I know that I shoud work in a minimal-photobleaching condition. The problem is that I am taking back a work that was started with a precise protocol and have to stick to it. Now I have to analyse my datas + former datas with this inconvient of photobleaching phenomenon.

In the meanwhile, we are looking into your solutions for futur recordings, espaccally since the analysis procedure is already too time consumming and fastidious.

If anyone has more suggestions I am happy to learn! :)

Looking forward

Dear Blandine,

If I understood right you are extracting the data first from ROIs and you are trying to fit them later to counter balance the bleaching of the sample?

Have you ever tried to correct the bleaching on the original image data with an exponential fitting in ImageJ/Fiji (might depend on the used installation) and to extract afterwards the numerical data from ROIs? We had very good results for Fura2 and Fluo4 with this approach.

Hello Sönke Weinert,

Yes this is what I am doing.

I should try that out, the "exponential with offset" plugin from Image J.

I will keep you inform.

Thank you,

Blandine