Replace _scans.tsv with _recordings.tsv

[tsalo]

Right now the _scans.tsv file keeps track of which acquisitions exist within a particular session. The word “scans” is a little domain-specific, especially as people begin to write BIDS formats for things like electrophysiology. Why not call this recordings, as this is more domain-general and still reflects the same overall idea?

Original authors: Unknown

[tsalo]

@jgrethe wrote:

Or perhaps _data.tsv which would be more domain agnostic

[robertoostenveld]

For me it is also not clear what precisely should go in the _scans.tsv file, i.e., what qualifies as data (or acquisitions) and what not.

In case of task-fMRI with events, a single line (with the bold.nii) would be added. In case of behavior only (no scanner involved), would the 'events.tsv` be added?

In case of EEG with recorded/digitized electrode positions, the eeg.vhdr file would be added but not the electrode positions ( I guess). But if only electrode positions are recorded and no EEG, as in this study and this corresponding data, would the electrodes.tsv be added?

[jbteves]

I would propose "acquisitions" though I will also advocate for #27, which would render this issue irrelevant.

[tsalo]

In Common Principles, acquisition refers to "a continuous uninterrupted block of time during which a brain scanning instrument was acquiring data according to particular scanning sequence/protocol." I actually think adopting "acquisitions" for the tsv file would be a great idea, @jbteves, independent of #27.

Here's my reasoning:

1. scans is currently confusing, as @robertoostenveld says. Acquisitions, on the other hand, are quite concrete. There may be edge cases, but generally speaking we know what an acquisition refers to.
2. In addition to different simultaneously acquired data types (e.g., fMRI + events), we have several entities that generally separate files from the same acquisition (e.g., echo). If, however, we have our (optional) tsv file dedicated to acquisitions, then we can pop out all of those pesky entities and get one acquisition time that applied to the whole file group, along with any other acquisition-specific information.
3. We could even perhaps include a column to provide acquisitions with identifiers, which could tie into https://github.com/bids-standard/bids-specification/issues/529. Basically, the "grouped scan collections" (scans from separate acquisitions that are parametrically linked and must be analyzed together) could be defined with this new "group_identifier" column in the "acquisitions" tsv.

[yarikoptic]

There is a need also very similar to

• https://github.com/bids-standard/bids-specification/issues/1771 with the need of temporal alignment of beh recording

so there _recordings is good,

With @neuromechanist we are working out stimuli/ formalization they have. With stimuli we have a need to align the stimuli whenever it is a video/audio covering entire run. But it is not collected data, and might not even be in that folder but rather just referenced from under stimuli/. But the point is that it also would relate to temporal alignment of various files pertinent to the experiment. If we decide somehow for it to be reflected in this file, then `_recordings.

[tsalo]

Recently, we started using descriptions.tsv to define values in desc entities in a dataset, and I have proposed extending that to other entities, including recording and acq (see https://github.com/bids-standard/bids-specification/issues/1706). I think the proposals in this issue would conflict with that.

[yarikoptic]

Good point, and 👍 on such generalization. Let's think alternative name