Why my Sholl analysis only works if scale is in pixels?

I am trying to perform sholl analysis in neurons, after applying a threshold and creating a mask (figure1). However, the analysis doesn’t take into account all the sampled values from the mask (figure2).
Then, I tried to apply the plugin Robust Automatic Threshold Selection (RATS), which allowed me to obtain the correct sholl mask (figure3) but all the values are relative to pixel unit. If before the analysis I try to set the scale again to µm, it doesn’t work anymore.

Anyone with a tip how about to overcome this problem?

Thanks in advance.

Welcome to the forum @Luciana_Fernandes,

(Hopefully I am understanding this correctly): The issue seems to be your initial is spatially calibrated, but the image created by the R.A.T.S plugin does not carry spatial calibration. The settings in the Sholl prompt are always specified in the units of the image, meaning that when changing scaling between images, you will have to scale the values in the Sholl prompt accordingly. A couple of tips:

  1. You can enable ‘Global Calibration’ in Image>Properties. This will impose the same spatial calibration among all images during a session (Note the ‘(G)’ on image titles
  2. You can specify a line ROI before calling the Sholl prompt. It will automatically set the center of analysis, and end radius
  3. If your images have only one cell and no surrounding objects, you can set ‘end radius’ to an exaggerated value. The plugin will simply ignore ‘out-of-image’ distances

Also, note the recent announcement that explains why the Sholl plugin has migrated to the Neuroanatomy update site.