How to count neuronal cells - proportional area - with DAB staining

Hi!
I’ve been trying to analyse the proportional area of NeuN staining (neuronal nuclei) in hippocampus slices (image).
I tried to follow the leaf example (https://imagej.nih.gov/ij/docs/pdfs/examples.pdf), but when I make the images binary, I lose information, because there are areas (delimited by the black line) in which there are a lot of particles, but since they are clustered, it doesn’t recognize as a lot of nuclei. I tried to use the plugin IHC toolbox, however, the quantification comes in number, and I need to know the stained area in relation to the selected area (each area has a number in the picture). Do you know if it’s possible with ImageJ?

I don’t know why the image doesn’t appear :disappointed:

Hello Natalia, try to save it as PNG and then upload it :slight_smile:

file >> saveas >> png

1 Like

Hello again Natalia, and welcome to the forum!

Regarding your question, first you say you want the proportional area but then it sounds like you want a count, not area. Then you say again that you need to know the stained area.

So, do you want a count AND area or just area? Do you want to know how much area of selection CA1 vs DG is stained? In my opinion you will not be able to reliably count these nuclei. The dentate gyrus and CA regions are extremely dense with cells, and especially the dentate gyrus granular cells. They are small and lie pretty much packed together.

How do you image these slices, and is this something you can change?

Maybe someone else on the forum has a clever way to do it. Personally I have never used the IHC tool.

If you upload a raw image without the markings for people to experiment with that would also help! It seems to confuse the rolling ball background subtraction, which might help you here.

In short, more information is needed. If you want area I’m sure it can be done. If you want count as well, I won’t promise anything!

Best,

Sverre

I’ve fixed this by doing exactly as @Sverre suggested…

1 Like

Thanks, @ehrenfeu!

Actually, this is a common issue caused by the forum allowing tiff images to be uploaded, but many browsers not supporting displaying them.
Post authors as well as forum moderators can edit these posts and perform the steps as linked below to embed png versions of these images:

Maybe we should create a FAQ entry somewhere to document this…

3 Likes

Thank you! :slight_smile:

Hi Sverre!

Thank you for your answer. I want the proportional area! I tryed the IHC tool because at first it seemed possible with it, since I could work with the color itself without having to make it binary. But you’re right, I totally agree with you, I could never count the nouclei in CA1 and DG.

What do you mean by “how do you image…”?

Here is the raw image…I saved as png

Thank you agan

Best,
Natalia

Ok, something like this? Still not sure what you mean by proportional area. Do you mean the relative size of the DG to CA?

I mean what is your microscope setup. But if you don’t want to count single nuclei, this is fine :slight_smile:

Yes! I need to “surround” the area! And by proportional area I mean: how many of the selected area is stained? So, the selected area would be 100% (CA1, for example) and the staining inside it would be <100% . The same for DG and other subareas… I won’t compare DG to CA1… I’m not sure if I made myself clear yet…

Ok, I think I know what you mean, not quite sure about your whole setup but if thats what you need. How do you set the same Area for the DG and CA in different slices? If you want to see the proportion of this area that is stained, then you need to define the CA and the DG at the same size in every slice… and of course different sections will have different hippocampal shapes…

This is what I did:

Image >> type >> 8-bit

Edit >> Invert

Process >> Subtract background, disable smoothing, 50 pixels.

Process >> Filters >> Gaussian Blur - Set value to 2.

Image >> Adjust >> Auto Threshold >> Triangle

And then select the regions you want. If you are doing this manually just use the wand tool, and add to manager. Click measure and get your area per region.

If you want to automate this procedure (very doable), I suggest you write a macro or script with the steps above, and use the analyze particles tool to select the regions.

HTH,

Sverre

Well, as you said I can’t set the same area for DG and CA1 in different slices, even because they get bigger as the section is more posterior… So I manually select the area, according to the mouse brain atlas…Regarding the setup, I acquired the images using a 4x objective, and I have the scale bar which I used to calibrate…

I tried this way you told me… it did work, thank you very very much!!

Best
Natalia

Ok Natalia, it sounds reasonable. As long as you account for the different sizes :slight_smile:

Glad it worked!

Informative post i really like it