I have been thinking about this question a bit now and wanted to ask the many folks in the NeuroStars community their from-the-hip, to-the-side, and otherwise off-the-cuff philosophical answers to the question…
I have seen “BIDS-ified” versions of bunches of MRI data in various states on various servers over the years now. However, I haven’t seen many BIDS-standard participants files or phenotype files which I have not had a hand in creating. Big projects like UK Biobank and the HCP data even, as far as I know, don’t bother with BIDS-standard participants or phenotype files.
Why is that? Is it because it feels like too heavy of a lift or does no one really want to do it? Is it because no one feels confident to do it or maybe feels the standards are changing too much or the participants and phenotype files are too young in the age of the standard? Is it because no one feels these files improve the value of a dataset? Or is there a more real possibility I’ve totally missed here?
Should I just make some effort to take data dictionaries of various big studies and output ready-to-go participants and phenotype files? Would that help anyone or am I just day-dreaming? If I’m just day-dreaming here, I think I need a better day-dream.
I’m as interested in hot takes as I am interested in thorough answers, so please drop some knowledge here. Thanks all!
Relevant BIDS standard sections: