I am currently looking into technologies I can use to process data for my project. I have good quality Spinal Cord data acquired ex-vivo from macaques (T2 + DWI). I heard a lot of good about SCT with human data, and followed closely the recent addition of ex-vivo support in the toolbox. I was wondering the SCT team was working on support for the NHP or had some ideas on how those data could be processed with the toolbox. Looking at the processing steps included in the tutorial pipelines, my hunch is that with good templates and atlases most steps should run out-of-the-box, but I would like to have your team’s opinions on that before starting to play with it.
Thank you for reaching out. It is definitely possible to use some of SCT’s function for non-human species. See for example the list of published work across species:
While the majority of publications are done with human data, some include non-humans, and notably one in NHP.
That being said, what is important for you to know is what kind of analysis you intend to do with your data. As you rightfully mentioned, a registration to the template requires to have a template. If you are thinking of a long term project where a template would be highly beneficial, we can consider creating a NHP spinal cord template, that could then be used with SCT. For example, we are currently building one for dogs, and one for human 7T data.
If you are “only” looking at segmenting the spinal cord, labeling vertebral levels, computing cross-sectional area measures along specific vertebral levels, extracting metrics (eg: MTR, diffusion FA, etc.) within the spinal cord mask, co-registering spinal cord across time points, etc. then you don’t need a template and can already use SCT to do all this.
Hope that helps
thank you very much for the fast response ! You are absolutely right that the analysis will drive most of the processing we’ll be doing. For now, the project is still at its beginnings, we are mostly trying to gather a set of tools to explore, have a good picture of what is done in the field.
Your answer validates most of my intuitions, which is great news ! I thought the template was more involved in the various processing steps, the fact it is not is really good. I’ll be looking further into SCT, thinker with it a bit.
We’ll probably look into template creation for the project down the line, but that will be for later. Thank you very much for the code you shared with me, it’ll be of great use for that !