ingest Jubrain ontology

[tgbugs]

v18 (has the current hierarchy of the brain areas stated in the "ontology" JSON):
Amunts, K., Eickhoff, S. B., Caspers, S., Bludau, S., & Mohlberg, H. (2019). Whole-brain parcellation of the JuBrain Cytoarchitectonic Atlas (v18) [Data set]. Human Brain Project Neuroinformatics Platform. https://doi.org/10.25493%2F8EGG-ZAR

v13 (only provides a list for the used probability maps (PMs)):
Amunts, K., Eickhoff, S. B., Caspers, S., Bludau, S., & Mohlberg, H. (2019). Whole-brain parcellation of the JuBrain Cytoarchitectonic Atlas (v13) [Data set]. Human Brain Project Neuroinformatics Platform. https://doi.org/10.25493%2FQ3ZS-NV6

[tgbugs]

High level regions are semantic, not based on data so using labelPartOf is probably reasonable. All high level regions lack a labelIndex, so may have to put them in readable and construct ids for them. Also issue with things like LB (Amygdala) which appears to have 4 different variants with different labels but that do not have sufficient information to differentiate them, and we may need to merge those back into a single term since the the variants are almost certainly MNI152 (nonlinear asymmetric 2009c) vs Colin27.