Count neurons in clusters

Dear all,
I am trying to figure out whether is possible to use cellprofiler to count neurons that grow in a cluster such as in the example image attached. In this case one staining is citoplasmic and the other one nuclear.
Do you have any suggestion about segmentation and pipeline?

Thanks for your help.


I took a quick stab at this (see Screen Shot). A few questions/thoughts:
(1) What do the channels mark? I assume the green channel is the neuronal marker? And the red?
(2) Clumped neurons are tough to analyze. One option that folks have used is to image all around the clump, where in the density is lower.
(3) Are there glia? If so (and indeed there look to be lots more nuclei than neurons), then it complicates using the traditional method of segmenting the nucleus -> Cell -> Cytoplasm.

Please note that commentary on the pipeline are written in the Module Notes at the top of many of the modules. This basically says:
(1) Split the color image
(2) Enhance the soma
(3) Remove the neurites
(4) Segment

I didn’t add anything for the red channel, since I didn’t know what it was. But in the Per_Image output table, you can get the counts that you seek.

Hope that helps,

DL_Count_Neurons.cppipe (8.88 KB)