- We now need a custom script to "claim" dicoms on BIC server before downloading them locally
- Some participants have multiple dicom dirs for same session (most likely from issues during scanning)
- Heuristic filter: select dicom_dir with most files
- Some dicoms are now "tarred" on bic server and need to be extracted before reorganizing
- Some dicoms do not have
visit
/session
info in the filename (i.e.MRI01
)- Cross-ref date with recruitment spreadsheet for identifying sessions
- Some dicoms have
derived
tag in the header- These are excluded from symlinking from
raw_dicom
todicom
before bids conversion
- These are excluded from symlinking from
dwi
error on volume count for all participants. See neurostars issuefmap
sidecars forPA
direction (e.g.sub-<>_ses-<>_acq-bold_dir-PA_run-1_epi.json
) have wrongPhaseEncodingDirection
for some participants. This will fail the downastream fmriprep processing at SDC stage.- List of participant ids:
PD00296
,PD00869
,PD01090
,PD01165
,PD01232
,PD01306
,PD01360
,PD01435
,PD01551
,PD01753
- This is manually changed from
i
-->j
- List of participant ids:
fmap
sidecars have wrongPhaseEncodingDirection
for some participants. This will cause failure of fmriprep processing at SDC stage.