Hi,
Changing any of the nmodes, cost function or skull stripping
should not change the relationship between the origsegs
and the firstseg as long as you have not altered the
left-right information in the nifti header. It is only the boundary
correction which differs between these two images. We did
have a bug a while ago which caused some left-right flipping
in certain cases, but that was fixed. Are you running the most
recent version? It is possible that this is the cause of your
problems, or that your skull stripping has altered the nifti
header information in a bad way.
I have not seen errors like the ones you noticed before.
http://www.fmrib.ox.ac.uk/cgi-bin/upload.cgi
If you cannot find any obvious cause of the problem and
you are definitely using the most recent version of FSL
then please upload your images to our server at:
and we'll have a look at them and try to track down
the problem. It would also be helpful if you could send
us your modified script.
All the best,
Mark
On 12 Mar 2010, at 22:13, Jason Stein wrote:
> Hi Brian,
>
> Thanks for your reply. Yes the all_fast_origsegs and
> all_fast_firstseg do
> not overlay. I ran
>
> run_first_all -b -i input.img -o output
>
> I did edit the run_first_all script to use different nmodes
> parameters than
> the defaults. I also edited the first_flirt to use normmi cost
> function
> because the default was not giving good registrations. I also used
> skull
> stripped brains for input because the non-skull stripped were not
> always
> providing good segmentations. Other than that though, no edits.
>
> Thanks for any advice,
> Jason
>
|