I have PAR/REC files for many participants. Everything works well for the BOLD acquisitions - in the sense that the correct pattern is found to run bidscoiner even if I just bidsmap one participant.
However, I have a problem with the MPRAGE. Here, the repetition_time changes for every subject, and I cannot find a way to define a regexp that would work.
As the original value - in the case you can see in the screenshot - is [9.929] I tried with the following:
[*] or *
\d+(\.\d+)? or \[\d+(\.\d+)?\]
However I get the same error when running bidscoiner:
ERROR | --> Skipping unknown run: [PAR] /data00/leonardo/bidscoin_test_advanced/data/PARREC/sub-020/ses-01/sub-020_ses-01_WIP_acq-MPRAGE_T1w_5_1.PAR
-> Re-run the bidsmapper and delete /data00/leonardo/bidscoin_test_advanced/data/bids/sub-020/ses-01 to solve this warning
I found out that I can just remove the corresponding line from the bidsmaps.yaml, so that would solve the issue for the bidscoiner (the value is still in the generated json), however it would be nice to have a solution with a regex and not to have to manually tweak the bidsmaps.yaml.
Thank you in advance for your help.
Leonardo
bidscoin 4.5.0 installed locally with pip in a venv.
The solution is to remove the repetition_time line from the template bidmap (i.e. bidsmap_dccn.yaml located in the .bidscoin configuration folder), only for the MPRAGE (assuming that the other sequences do not change their TR). By doing it in the template you are fixing it for once and for all.
p.s. The bidsmap does not list subjects, nor is it done per subject. It lists the different scan types, doesn’t matter how many you have of each type, they are all considered identical. Otherwise editing would become an awful lot of work (as then the list would grow with the nr of subjects).
Thank you very much Marcel!
I hardly have experience with Philips data, is it common for the TR of pulse sequences to change (in the minor digits) from acquisition to acquisition? If so, then I will change the PAR/REC section of the template bidsmap for everyone
Hello Marcel,
I asked to an MR Physicist friend of mine who has a lot of experience working with Philips scanners and developing sequences (Jeroen Siero at UMC Utrecht). He told me that Philips has an option for setting the TR as shortest - which indeed I have seen used many times - and in that case, depending on the angulation of the FOV, TR can vary across subjects/acquisitions.
Thanks, that’s very helpful. For the upcoming release I will remove repetition_time from the template (fortunately there are enough other parameters left to make the scan type unique)
Thank you Marcel! I believe that it will be very appreciated from Philips users.
P.S. Bidscoin is a really great piece of software!