I read in the FreeSurfer wiki that using T2 or FLAIR images could improve the skull-stripping giving a better outcome for pial surface reconstruction. Did anyone actually have tried this? is worth the effort?
I'm having trouble running the command, because or FS don't recognize the -FLAIR or -FLAIRpial flags, or can't be able to find the image I specified the path to. I'm using FreeSurfer v.5.1.0.
Since the last thing happened to me when I first tried the recon-all command, and the FS tutorial don't specify that you have to call your original MPRAGE-3D as 001.mgz and save it on the orig/mri path, could this happen with the FLAIR image as well? FLAIR image has to be in a particular folder or named specifically?
Thank you all,
J
Hi Jonatan,
Firstly, version 5.1.0 is outdated and very old. You should be using the latest version (version 6.0).
To use the FLAIR (in addition to T1) for recon-all, the instructions are here: https://surfer.nmr.mgh.harvard.edu/fswiki/recon-all . Look under the heading "Using T2 or FLAIR data to improve pial surfaces" where the instructions are clearly described.
Regards,
Jerome
thanks Jerome,
since FreeSurfer version 5.1.0 is from 2011 (never really thought about this!), maybe this command is not allowed in older versions than the current 6.0
We started using a FLAIR sequence in our scans, and the -FLAIRpial flag dramatically reduced the errors. I recommend everyone collect a FLAIR sequence (or at least a T2 sequence) if you want accurate cortical surfaces without a lot of time spent making manual corrections. As the wiki says, the input scans can be specified with the "-FLAIR " (or -T2) flag, similarly to the "-i " flag, along with the "-FLAIRpial" (or "-T2pial") flag that turns on the surface correction.
Does anyone know if "T2"-images are supposed to be T2-weighted or actual T2 maps?
But quantitative T2 maps will work better? Does Freesurfer use quantitative priors? What about T2*-weighted images/q maps?
Hi dears
Which of them is better to improve pial surfaces in Freesurfer?
T2 or FLAIR data?
I have tried both T2 and FLAIR. For me FLAIR is much better. I think it is because of the image resolution, which is .5 x .5 x 3 mm3 for T2 and 1 x 1 x 1.9 mm3 for FLAIR. I recommend you to try both of them at first to decide which one is better for you. However, when the pial surface is much improved for my older MRI scans, it becomes much worse for many of my MRI scans acquired since 2018 (but not all of them, intriguingly). This is caused by an error in co-registering the FLAIR with T1. I tried to use FLAIR images that have been coregistered with T1 using SPM 12, but it does not solve the problem. We are using Siemens 3T Trio for collecting MRI scans. Does anyone have any clues why this could happen?
I did a search in the Freesurfer mail list and found that this feature has not been tested extensively in 6.0: https://www.mail-archive.com/[email protected]/msg51413.html. So we need to wait for the next version.
I do believe that to be useful FLAIR has to be 3D. Anyone tried with 2D-FLAIR?
- Badges
- Science topic
- Similar topics
- Philosophy
- Moral Philosophy
- Free Will