Since c-fos, is also activated because of stress, it is not a simple quantification. If you are comparing two groups like treated and untreated, if you can separate the amygdala alone, it would be interesting if you can do western blot for c-fos, that would enhance your data. As you mentioned that it varies from animal to animal, it would be good to cut the whole amygdala and stain every other sections or every sections for c-fos and then count the positive cells as well the intensity quantification also would help to compare or to report the stand alone data..May be one option, use some voxel based quantification will reduce the ambiguity in scaling up.

Thank-you for your answer - I am looking at in relation to stress, and have 3 groups (naive/sham/test). Unfortunately the brain tissue is fixed (and sectioned), so if I wanted to do a western, I imagine I'd have to run the whole things again and harvest fresh tissue.

Currently, I count positive cells and divide by the count area (i.e. CeM/CeL/CeC).

I've looked into automation using ImageJ, but my images are not consistent enough for this to be reliable (at least with my limited abilities with Image J).

It is common to use cell count for IHC results. I agree with the previous person that it is a good idea to pick sections that have been sliced in the same plane. I typically slice through the whole amygdala and use every third slice for Immunos. I also check out all of the slices before starting an immuno to try to make sure that across each animal and group they are roughly the same plane. This should help take some of the variability out of your counts. ImageJ is great for this and I typically find the slice with the darkest staining and set thresholds based on that image that I then keep constant for analyzing all the other images.

I think cfos immunos are pretty standard so I disagree somewhat with the previous comment. If you can get your immuno standardized and are able to show convincing representative images in your publication I'm sure that there's a reason to do a Western as well.

Hope that helps, or gives you a few things to think about (if you're still working on this, or if you revisit immunos in the future).