Hi,
There seems to be plenty of contrast in these spin echo images.
I suspect the problems are coming from the large FOV and/or poor brain extraction of the structural images. We recommend using fslreorient2std and robustfov (or fslroi with manual selection) to provide close alignment and FOV of the structural images. I think if you did this, and checked the brain extraction results (which will probably be improved by this) then this should fix the problems.
All the best,
Mark
> On 18 May 2017, at 15:19, Mark Pinsk <[log in to unmask]> wrote:
>
> Oh, I just spotted one mistake: the effective echo spacing didn't account for GRAPPA=2.
>
> So that fixed the bad output in many subjects, except in those subjects where the magnitude to highres registration failed.
>
> Any idea why that registration is failing us? Maybe the spin echo image contrast too poor (using 8000ms TR / 66ms TE) ?
|