That all sounds right, thanks a lot Philipp!
Juan Eugenio Iglesias
Senior research fellow
CMIC (UCL), MGH (HMS) and CSAIL (MIT)
From: <freesurfer-bounces@nmr.mgh.harvard.edu> on behalf of "Saemann, Philipp" <saemann@psych.mpg.de>
Reply-To: Freesurfer support list <freesurfer@nmr.mgh.harvard.edu>
Date: Friday, April 2, 2021 at 03:51
To: Freesurfer support list <freesurfer@nmr.mgh.harvard.edu>, Umesh Rurdrapatna <umeshrs@gmail.com>
Subject: Re: [Freesurfer] 2mm thick image for hippocampal segmentation.
External Email - Use Caution
Hi,
I try to answer but Eugenio may also add comments. Several points that are important:
- 1.5 Tesla and 1x1x1 mm3 are at the limit of what the algorithm can manage.
- Post-aquisition reslicing to a higher resolution does not really increase the information content. Internally the algorithm works at a much higher resolution, anyway.
- If possible, switch to 3 Tesla.
- If possible and this is a prospective study, add a coronal T2 image and use the bi-channel option. See papers of Iglesias et al. on this.
- What you could also do on a 1.5 T is acquire the T1WI double and average it after coregistration, this increases the SNR.
- Generally, isometric resolutions are good due to the symmetry, of course.
- Also on 1.5 T, you may go bit below 1 mm with good head immobilization, and a longer scanning time. BRAVO in general is a good T1 option.
Best,
Philipp
Von: freesurfer-bounces@nmr.mgh.harvard.edu <freesurfer-bounces@nmr.mgh.harvard.edu> im Auftrag von Ramesh Babu <mgrameshbabu2013@gmail.com>
Gesendet: Freitag, 2. April 2021 09:29
An: Freesurfer support list; Umesh Rurdrapatna
Betreff: Re: [Freesurfer] 2mm thick image for hippocampal segmentation.
External Email - Use Caution
Dear experts,
Sorry for reposting. I am expecting some response from experts, so that I can proceed with confidence.
For our project work, we are looking for hippocampal and its subfield volumes. Using GE 1.5T scanner, we tried a 3D BRAVO sequence with 1x1x1 resolution, 256x256x160
dimension with 8* flip angle which produced images with a lot of noise. Then we tried the same sequence with 1x1x2 resolution, and 256x256x80 FOV, which produced a clean image without noise. When we reformatted the images, it produces 0.5x0.5x1 resolution
with 512x512x172 dimension.
Is this sequence good for hippocampal volume estimation and subfield segmentation (also for whole-brain VBM analysis)?
Could you please suggest to me a correct sequence, if need to be changed?
Regards
Ramesh