Dear Tyler,
Quite often the fieldmaps show very noisy voxels around the
edge of the brain due to partial volume - which they are very
sensitive to. This is why there is a more aggressive masking
done to the fieldmap. However, this does not mask the brain,
as the fieldmap is extrapolated at the edges, which means that
the parts of the brain outside this fieldmap mask get shifted
consistently and smoothly, avoiding the jagged, rough shifts
induced by noisy voxels.
The subtraction of the median value, as calculated by
fslstats, is to avoid global shifting of the image - which can
sometimes be very large and is completely arbitrary and
unnecessary.
I hope this explains things for you.
All the best,
Mark
On 28 Feb 2011, at 21:18, Tyler Blazey wrote:
> Hi,
>
> I have a few questions about how the FEAT GUI uses prelude and fugue for
> field map correction First, looking at the log file, the script the GUI
> launches runs several commands to create a more aggressive mask than is
> inputed. How necessary is this? I copied all the commands in the log
> excluding the masking commands, and couldn't notice much of a difference.
>
> Second, at one point the script launches fslstats ( fslstats FM_UD_fmap -k
> FM_UD_fmap_mag_brain_mask -P 50 ), and then subtracts the result from the
> field map. What exactly is the purpose of this?
>
> Thanks in advance for any help,
>
> - Tyler Blazey
>
|