Is there any way to include video of subjects? It seems it may be possible to sneak it in under tasks but that does not seem correct. We have video of subjects both in the bore and eye vide (sometimes their eyes are closed, and our eye tracker cannot track them, but it is possible to apply custom algorithms to get meaningful data despite the eye being closed).
I’m not really sure on this. Searching through the standard and the mailing list doesn’t turn up much. Video stimuli would generally go in the /stimuli subdirectory, so that’s at least a valid option for now, even if it’s semantically a little off.
@satra @mgxd I know you have audio recordings of subjects. Where do you store them?
we keep it in sourcedata/sub-<>/audio - though I’m not sure how compliant this is
Under a video/ heading or also in sourcedata/sub-<>/stimuli?
in our case it’s audio/
sourcedata/ is free-form, as it can contain any data that is not yet translated into BIDS format.
So, to be clear, we have video data of every run for each of our subjects. There is a part of the standard, at least as originally published in Nature, that seems to suggest that it is permissible to place non-BIDS-compliant data in the standard directories wherever common sense would suggest it would go, and hope a standard will come out some day. Is that correct? "For example one may want to include eye tracking data in a vendor specific format that is not covered by this standard. The most sensible place to put it is next to the continuous recording file with the same naming scheme but different extensions. "
I think it appears that in practice putting the video files into sourcedata/ perhaps may be a strong place. So the proposal would be putting the video data into sourcedata/sub-<>/video directory. You may follow the general naming convention of BIDS to more easily query the data and match your video files with the imaging. As the specification continues to extend and cover video data, the files can be merged over to the main BIDS directory.