Image J, threshold question

[Jordan GUMC]

Hi,
I was wondering if someone could help me with Image J, specifically with the thresholding  function.

We have a multitude of brain sections that were stained with a BCIP/NBT kit, resulting in a purple precipitate in cells/processes showing increased oxidative stress. We are trying to calculate the area of the cells stained using the kit.

The problem is that even though all of the slides were done as a cohort, in the same staining dish, some of the sections have a much darker background than others. Most sections are almost completely white (clear) with nicely outlined cells/processes stained, however others (Cells) are surrounded by a deeper purple stained background.

We have tried various threshold values, it seems 120 (out of the 255) seems to work the best, however on some of the darker stained sections, the background is coming through intensely, and I know just by looking that the measurement value is not correct. We even tried taking some background measurements and subtracting them, but picking the background areas is not scientifically appropriate due to the bias.

I have searched for various ways to decrease the background without decreasing the positively stained cells/processes, but all of the ways (Shading, contrast, binary) don't seem to work.

Any thoughts on what can be done? We need to keep the threshold the same through every slide to maintain a constant value, I assume.

Thank you,
Jordan
Post-doc, Dept. Neuroscience, Georgetown