khanlab / hippunfold

BIDS App for Hippunfold (automated hippocampal unfolding and subfield segmentation)
https://hippunfold.readthedocs.io
MIT License
47 stars 12 forks source link

expanded DG #298

Closed jordandekraker closed 2 weeks ago

jordandekraker commented 2 weeks ago

Currently HippUnfold generates only a midthickness surface of the DG (no inner/outer, and no morphometry). In typical MRI this makes sense because the DG is only around 300um thick, and so multiple surfaces would be meaningless (and messy because of missing IO laplace coords). However, in some cases (Allen mouse, BigBrain) it may make sense to produce these outputs as well.

Currently the label wildcard is manually set for a few rules in https://github.com/khanlab/hippunfold/blob/master/hippunfold/workflow/rules/gifti.smk. We could consider moving those constraints to the config section instead (exposing this is generally good for transparency), and we could add an optional flag --procDG

The flag could work as following:

  --procDG:
    choices:
      - 'none'
      - 'limited'
      - 'full'
    default: 'limited'

Where limited is the current behaviour. A config would then specify a list of which label to run for each rule for each choice.

This is probably not high priority now, but may be useful if/when more 3D microscale data becomes available.

jordandekraker commented 2 weeks ago

With https://github.com/khanlab/hippunfold/pull/300/commits/92113225b9ea4cbb6353e2651fe0b66af54080f8, this is now more accessible (though it will still require editing the config instead of a CLI argument