We are encountering some issues with the brain masks and (temporal/anatomical) CompCor ROIs in some of our subjects’ multiband, multiecho data processed usingfMRIPrep version 1.5.1rc2 (see pictures below). Their brain mask and brain tissue segmentation of the T1w outputs appear normal. One of the subjects has enlarged ventricles, but we do not note anything anatomically abnormal in the other subject. In addition to the images below, this is a link to a Google Drive folder containing the compressed fMRIPrep outputs for the 2 problematic subjects. We would appreciate any guidance on how to proceed with these subjects. Thanks in advance!
Note that the masks appear to have rotated axes relative to the original file. I suspect there’s an affine validation/conformation step gone wrong. Could you share a listing of an affected subject? We’ll also need to see the qform/sform matrices and codes for all relevant BOLD/SBREF files.
Hi @effigies, we have included the raw data for the 2 affected subjects in a sub-folder within the original Google Drive folder, in addition to sub-folders containing the qform/sform matrices and codes for all relevant BOLD/SBREF files. For reference, those files relate to the following runs for each subject, and are named accordingly:
Sub559
- Arrows run 1
- Arrows run 2
- Arrows run 3
- Encoding run 1
- Retrieval run 1
Sub637
- Arrows run 2
- Arrows run 3
- Encoding run 2
- Localizer run 1
We are also encountering some brain mask issues with another subject, which we suspect is due to the intensity non-uniformity (INU) correction of the T1w (as seen by comparing the original image (Raw Data/sub-635/anat/sub-635_acq-highres_run-01_T1w.nii.gz) with the INU corrected image (sub-635/anat/sub-635_desc-preproc_T1w.nii.gz) at this Google Drive link). We experienced this problem with a different subject, and after re-extracting their data without defacing it, we were able to resolve their mask issues. However, this fix did not work for the subject pictured below. Again, we really appreciate any guidance provided regarding how to proceed! Thanks so much!
Hi everyone, we have not been able to resolve the issues described above regarding the brain masks and (temporal/anatomical) CompCor ROIs in some of our subjects’ multiband, multiecho data processed usingfMRIPrep version 1.5.1rc2. If anyone can offer any advice on how to proceed, it would be much appreciated! Thanks so much!
Hi @cphaneuf, it looks like I experience the same problem as you regarding the brain masks and (temporal/anatomical) CompCor ROIs, also on multiband, multiecho data. Have you been able to solve this issue over time?
Hi @fabienne, we re-processed our data using a later version of fMRIPrep (20.0.6). While this did not resolve all of the brain mask issues for our problematic subjects (Sub559 and Sub637, above), the outputs did improve substantially.
1 Like