I am wondering how people organize the data for sleep-stage scoring when they have polysomnigraphy data.
Looking for some “precedent” on openneuro I can find several options already.
There is a dataset that has 2 events file but one has an acq-scoring entity: OpenNeuro
This one has the scores in the sourcedata: OpenNeuro
For this dataset we don’t know AFAICT: OpenNeuro
Any suggestions? Or other examples out there?
I guess this is the same thing for graphical element annotated from the EEG such as spindles, slow-wave or epileptic activity.
So far, I always merged all the events file since the BIDS recommendation are saying " A single event file MAY include any combination of stimulus, response, and other events" [1].
To me, it implies that we cannot have multiple events files. Maybe the difference between the different event can be explained in the side-card json file
yup that would have been my first approach as well but I am curious to see how people tackled this.
There is an option to include HED tags in BIDS that describe sleep stages in a machine readable way. You can use the tags from the HED-SCORE schema library, see here: https://www.hedtags.org/display_hed.html
For some examples on HED-tags in BIDS, here is a pull request in progress:
MANAGED BY INCF
