I am trying to use the AVA_V3 atlas downloaded from here: https://scalablebrainatlas.incf.org/mouse/ABA_v3
to coregister region labels onto my own 3D mouse brain data. I did download the annotated .nii atlas and noticed region labels were denoted by differing grayscale values, which I presume are translated to actual names using the label index mapper?: Map label indices to region names.
My first question would be how the names in the label mapper match up to the grayscale value? I presume in the .json output list from the label mapper, that the index value of each name corresponds to the grayscale value in the image in ascending order? So a grayscale value of 0 would denote “root”, and 1 would denote “Basic cell groups and regions”, and 2 would denote “Cerebrum”, and so forth.
However, this brings me to next question. I noticed there are a huge number of labels in here, so I was wondering how gross brain region labels such as “cerebral cortex” would correspond to the actual atlas. For example, in the .json labels list, “cerebral cortex” is the fourth indexed label. However for dozens of labels afterwards, there appears to be subregions of “cerebral cortex” labelled, so if I wanted to actually have a segmented “cerebral cortex” only, would I have to make a python script that found all the subregions in “cerebral cortex” in the atlas and created an superregion combined from all of them, for example? Or is there another atlas I should be using, that labels the gross regions (such as cerebral cortex, hippocampus, caudate, etc) of the mouse brain instead of the hundreds of subregions?
Thanks for any help to anyone who is familiar working with these labels. I really appreciate it!