**Attached is the conversation, where it left off in August, some reason the thread did not get appended when I replied.
Dear Andreas,
Please forgive me for the extreme delay in response...
The original data from the scanner looks similar. After posting, my initial thought was that this was an isolated artifact in the b=0 image I was using as a reference - and that the subsequent nodif_brain_mask file was eliminating any FA signal down the analysis pipeline. However, it seems as if every slice in my raw data set (both b=0 and b=1000) includes this artifact.
DP
>Hi,
>how does the original dat look like?
>Cheers,
>Andreas
Am 22.08.13 23:39 schrieb "David Paul" unter <[log in to unmask]>:
>Dear All,
>
>I recently encountered an artifact in one of my datasets (see image
>attached to link) in which a portion of the parieto-occipital lobe is
>missing (i.e. void). All of the orientation directions are accurate in
>the major tracts. It does not seem to be an issue with inadequate
>scanning volume, and the MPRAGE sequence was fine. All other subjects
>scanned on this protocol displayed no such artifacts (SS-EPI; 60
>diffusion directions, TR/TE=8900/86 ms, b = 1000 ms/mm^2, 70 slices,
>2x2x2 mm resolution).
>
>https://docs.google.com/file/d/0B-nHo9UdMIi8NTYtSnVQYW1WYlU/edit?usp=shari
>ng
>
>On repeat scanning, the same artifact in the same exact location appeared
>again - I highly doubt this is eddy current/motion, and it does not look
>typical of EPI distortions (which I have corrected for using fsl).
>
>1. Has anybody had any experience with these types of "void' artifacts,
>and how do you deal with them?
>2. Does this artifact preclude one from reliably measuring DTI-indices in
>say the genu of the corpus callosum?
>
>Any help would be greatly appreciated.
|