Open effigies opened 4 years ago
It violates a bit the notion of "raw" datasets as data-off-the-scanner, but if we consider skull-stripping as a kind of extreme defacing (which is how it seems to have been used), then it's not inconsistent with previous decisions.
This argument sounds reasonable to me.
I'll go ahead and draft a PR, and people can argue there.
Derivatives has introduced a REQUIRED metadata field
SkullStripped
to imaging files.https://bids-specification.readthedocs.io/en/stable/05-derivatives/03-imaging.html#preprocessed-coregistered-andor-resampled-volumes
One situation we have in OpenNeuro is that some datasets were skull-stripped instead of merely defaced. This has led to fMRIPrep developing a heuristic to detect skull-stripped images. If datasets could specify in metadata that images have been skullstripped, it would bypass the need to continually refine the heuristic to account for false positives and negatives.
It violates a bit the notion of "raw" datasets as data-off-the-scanner, but if we consider skull-stripping as a kind of extreme defacing (which is how it seems to have been used), then it's not inconsistent with previous decisions.
I would like to make this an OPTIONAL field for raw MR data. How do people feel about this?