Hello NeuroStars-Community, i am quite new to fMRI preprocessing, and i got a dataset from a few years ago to analyse. Summary of what happened: The preprocessing was already run once using fmriprep v.21.0.2. But the output looks quite strange and while inspecting the raw BIDS, i found that espcia…

Hi @JanSch and welcome to neurostars! Do results improve when processing with the newest version (23.2.1 at this time)? Or alternatively could —use-syn-sdc and ignoring the dedicated field maps work for you? Best, Steven

Hi @Steven , thank you for your reply. Using version 23.2.1 did improve the results from the sdc partially: Old “Correction”, where the scan is pushed in a pyramid like shape [correctedSDC] vs. the new Correction, which looks much better, but still the frontal areas do not look very good [cor…

Hi @Steven , i tried running fmriprep with and --ignore fieldmaps at first seems to provide good results: Results for one task with functional scans: [Bildschirmfoto 2024-03-18 um 17.38.20] But for another task the results look different: [Bildschirmfoto 2024-03-18 um 17.38.55] When using the…

Hi @JanSch , [image] JanSch: Icould so far not get to run --use-sync-sdc, where i get an error, which i was so far not able to work around. What was the error?

Hi @Steven , here is the error report: Traceback (most recent call last): File "/opt/conda/envs/fmriprep/lib/python3.10/site-packages/nipype/pipeline/plugins/multiproc.py", line 67, in run_node result["result"] = node.run(updatehash=updatehash) File "/opt/conda/envs/fmriprep/lib/python3.10/…

Hi @JanSch , [image] JanSch: Stderr: Killed This means you are out of memory. Upping your resources should get you past this. Best, Steven

@Steven Ah, thank you for the help. I increased it and started running again. I will update here if it finished!

Hi @Steven , i upped the memory to 30 GB, but still encountered the same error. In the next few weeks i am unfortunately not able to further increase memory, because we would need to change it on the server directly. As soon as i am able to devote more memory i will try again. @jsein thank you for …

Hey Neurostars Community, as an update: we did get synBOLD-disco running which provided really good results so far. The Option syn-sdc worked interestingly good for our first test subject, however in the QA the results for the alignment of the fieldmap and the EPI are clearly faulty/buggy: BOLD R…

Fieldmaps with wrap around still usable?

jsein March 19, 2024, 3:31pm 9

Hi @JanSch, what you could try is to use Synbold-DISCO, if you have a T1w image acquired on the same subject.

There is a discussion on how to use this tool methodology with fmriprep: