This might cause the fiber tracking issue, but would not result in noisy
data. How do the raw data look in terms of noise, as compared to your other
data?
Peace,
Matt.
-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf
Of Martin Kavec
Sent: Sunday, April 05, 2009 2:28 PM
To: [log in to unmask]
Subject: Re: [FSL] What is the Problem with that DTI data?
Hi Markus,
I work with Philips Achieva and have good experiences with it. I always use
bvecs and bvals produces by MRIcron, when I do dicom to nifti conversion. I
would guess that the physicist gave you a one single gradient table which I
suppose you use as is. The fact is that the table has to be corrected
(rotated) for imaging slice angulation. MRIcron does it automatically for
you.
So I would suggest you to use MRIcron, and its simple GUI, dcm2nii to
convert
from dicom to nifti, and use also bvecs and bvals. I am almost sure that you
will get correct results.
Best,
Martin
On Sunday 05 April 2009 20:58:55 Markus Gschwind wrote:
> Dear DTI experts!
>
> For some month now, I am doing successful FDTing, mostly with Data from
> several Siemens Trio.
>
> Here I got data from a Philips Achiva. Unlike with data from the Siemens
> scanner, here the bvecs and bvals are not recreated by MRIConvert. But I
> got them from the Physicist.
>
> However, I have the feeling that that data are not good, but I am unable
to
> say in what.
> In a previous message Saad described how to check for not correct bvecs
and
> to correct them.
>
> I have not found that these bvecs are really grossly wrong (correct
colors,
> more or less nearly corect lines). I also tried to invert the signs, but
it
> was not visible to me if there was an improvement.
>
> However the dti_FA has a lot more noisy high intensity voxels and
> fibertrackging looks quite strange (see attachment).
>
> Would someone mind having a short look at that data? Could I upload them?
>
> Thank you so much!
> Markus
|