[RF] Review BIDS compliance of outputs

[tsalo]

Summary

At a glance, all of the functional derivatives seem to have a bold suffix- even things like time series TSV files.

Additional details

While functional derivatives aren't currently well supported in the main specification, BEP012 has been pretty stable for a while now, and I believe it's been adopted by fMRIPrep. We can refer to bids-standard/bids-specification#519 for inspiration.

[tsalo]

Just a first pass:

• _desc-timeseries_bold.tsv --> _timeseries.tsv
  • Should have column headers and a sidecar file explaining the columns.
• _desc-connectivity_bold.tsv --> _measure-pearsoncorrelation_conmat.tsv or something similar. Based on BEP017.
  • Also may require metadata fields to label rows and columns, or _nodelabels file(s).
• _desc-framewisedisplacement_bold.tsv --> _desc-framewisedisplacement_motion.tsv
• _tmask.tsv --> _outliers.tsv
• _desc-alff_bold.nii.gz --> _alff.nii.gz
• _desc-alff_smooth_bold.nii.gz --> _desc-smooth_alff.nii.gz
• _desc-reho_bold.nii.gz --> _reho.nii.gz
• _desc-reho_smooth_bold.nii.gz --> _desc-smooth_reho.nii.gz
• _desc-residual_bold.nii.gz --> _desc-denoised_bold.nii.gz
• _desc-residual_smooth_bold.nii.gz --> _desc-denoisedSmoothed_bold.nii.gz
• _desc-qc_bold.csv --> This isn't BIDS-ish. I'd recommend putting these metrics in the denoised BOLD file's JSON, like MRIQC does. Then analysis_level = "group" would collect the outputs into a TSV file at the top of the derivative directory.

[mattcieslak]

possibly make a session-level TSV that has all the metrics. Maybe scans.tsv. Maybe roadmap for now

[kahinimehta]

Makes sense! Also:

• make the *conn.nii --> to .tsv for cifti correlations