Automatic cell counting in brain slices


I’m new to immunohistochemistry, and I’ve been trying to develop an automated counting system for a group of interneurons within the mouse striatum. In short, I would like to be able to identify the number of neurons within a slice and obtain their xy coordinates.

I have tried a number of online methods on the images I have obtained but none appear to satisfactorily identify all the neurons, or they appear to highlight some of the processes that are also highly stained. Part of the issue is the high background noise that I’ve gotten with the antibody. I’m not very sure what to settings to adjust to make some of the methods found online to work, so any help would be very much appreciated. I would be happy to provide an example image.

@ohsoel The best way for us to help you is to share an original image file with us… you can just select a single, representative slice to share to start. And indicate/annotate in another snapshot where the cells are that you want to segment… :slight_smile:

Thanks for your help. I was not able to upload the picture to this site, but I have tagged a dropbox link below.

The first image is the original image file, and the second is an example close up of the neurons that I want to be able to automatically count.

Thanks again! :slight_smile:

Using some pre-processing filters like Gaussian blur followed by Maximum should help a bit… and setting an auto-threshold using MaxEntropy seems to then do a decent job. Give those steps a try first.

run("Gaussian Blur...", "sigma=4");
run("Maximum...", "radius=4");
setAutoThreshold("MaxEntropy dark");
run("Create Mask");
run("Analyze Particles...", "display exclude add");

For now… you should look at the following links to get you started on segmenting those cells:

This section of links will help you when you need/want to automate things:

1 Like