 |
 |
Dear FSL mailinglist, After applying tensor fitting on my diffusion data (either with DTIFIT or in-house scripts), I noticed a crown/ring of white (FA>1 !) voxels around the brain in the resulting FA image, which is obviously not anatomically valid (and present in all my datasets).. At first I thought this was due 'leftover' dura/csf after the use of a 'too large' mask image based on a structural sequence (MP2RAGE), which is a hell of a job to skullstrip (tips on that are very welcome by the way!). However, using DTIFIT, with a brain mask based on the b0 scan, instead of the high resolution structural scan, I get this problem too. Could this problem be caused by a sequence effect / scanning artefact, which I need to correct for? Or is this a known artefact in tensor fitting that I can control for? In a previous study I used a different diffusion sequence, and did not face this problem. Currently I have a dataset that was collected with a multiband sequence (cmrr). I can hardly imagine, but maybe the tensor fitting calculations I use are not optimal for this sequence? I am sure I need to get rid of this ring artefact somehow, and should not ignore it, since I will use the FA images for TBSS and image registration. I was thinking to 'simply' get dir of these ring voxels by thresholding, so that all FA>1 voxels are removed (within the brain no voxel exceeds FA=1), or by first creating a ring 'mask', and then subtracting that from the FA image.. does that sound logical, or do I then corrupt the data in a way I cannot validly use it anymore for the set purposes?? I hope someone can shed some light on the issue.. Thank you in advance! Zsuzsi ----- Z. Sjoerds, PhD Max Planck Institute for Human Cognitive and Brain Sciences Fellow-Group Cognitive and Affective Control of Behavioral Adaptation Group Schlagenhauf, room C211 Stephanstraβe 1A 04103 Leipzig Germany [T]: +49 (0) 341 9940 2471 [E]: [log in to unmask] / [log in to unmask]