Dear Pedro,
I think it is very unlikely that your sequence has a TR of 3100 seconds (or 51 minutes!)
and it is much more likely to be 3100 milliseconds (or 3.1 seconds!). The TR that is
reported for such a sequence should be the trslc, and the TR value for use with pulse
needs to be trslc * ny * numslc (so, multiply by 64 in your case).
When I did the following:
pulse -i /usr/local/fsl/data/possum/brain.nii.gz -o pulse --te=0.098686 --tr=198.4 --trslc=3.100 --nx=64 --ny=64 --dx=0.427 --dy=0.427 --maxG=0.055 --riset=0.00022 --bw=100000 --numvol=1 --numslc=1 --slcthk=0.0025 --zstart=0.09 --seq=ge --slcdir=z+ --readdir=x+ --phasedir=y+ --gap=0.0 -v --cover=100 --angle=90
It worked fine for me. I'm not sure if it will generate a good image - I haven't
tested that as it will take longer, but it was fine for generating the pulse sequence.
As to your alternative method - that would be fine if you don't want to do anything
much with the structural image (e.g. have motion, or look at modifications of the
pulse sequence parameters). If you just want a single, static structural image then
you can use BrainWeb, or just combine the brain input images in the possum
data directory with appropriate weightings (as this is all that BrainWeb does).
All the best,
Mark
On 3 May 2011, at 09:47, Pedro Mendes wrote:
> Dear Mark,
>
> As I understand tr must be equal to the trslc*numslc. As I want just one slice so the times must be the same.
> About the times there was some confusing, so I found in an article sometimes used in a really experiment:
> TE=98.686s
> TR=3100s
> After find these times I run the 'pulse' and the same error occur. The command-line that I used was the following:
>
> pulse -i brain.nii.gz -o pulse --te=98.686 --tr=3100 --trslc=3100 --nx=64 --ny=64 --dx=0.427 --dy=0.427 --maxG=0.055 --riset=0.00022 --bw=100000 --numvol=1 --numslc=1 --slcthk=0.0025 --zstart=0.09 --seq=ge --slcdir=z+ --readdir=x+ --phasedir=y+ --gap=0.0 -v --cover=100 --angle=90
>
>
> Even this issue is interesting for me and my thesis, I probable find another solution for my simulation problem, that is simulate anatomical images directly from BrainWeb, and used possum to generate EPI images, where I should used the same object that BrainWeb.
> I do not know if this is the better way to solve the problem, but I would like to ear your opinion.
>
> Best Regards,
>
> Pedro Mendes
>
>
|