Hi,
 
----- Original Message -----
From: [log in to unmask] href="mailto:[log in to unmask]">Yuhong Liu
To: [log in to unmask] href="mailto:[log in to unmask]">[log in to unmask]
Sent: Saturday, October 02, 2010 2:40 AM
Subject: [FSL] dtifit output

Dear FSL experts:

I am new fsl users, I have some questions.
I  generated FA image by following the following steps
1: eddy_correct
2: Bet, threshod 0.5
3: dtifit


In the generated FA image,  there are very high intensity in the brain boundary,  and there are some voxels FA > 1, see the attached figure.  The FA image seems to look unclear.

 
The high FA intensity in the brain boundary is caused by noise, as well as by residual misalignment between the DW volumes that has not been corrected by eddy_correct. Regarding the FA>1 voxels, this is caused by negative eigenvalues estimates, due to noise artifacts. There are many previous post on this issue, please take a look at the archives.


I used DWIs of the 32 directions to generate FA.  I converted decom to .img (.hdr) by MRIConvert.  There are 34 volumes in this 4D image, and I checked bvecs and bvals (see the attached file) and found:
the last gradient direction in bvecs  is 0 0 0, and the last value in bvals is 800,  I am not sure that these files are correct.But I delete the last volume in 4D image, remove the last gradient direction (0 0 0 ) in bvecs and the last b = 800 in bvals, but I got errors in ditfit:
Errors:  bvecs and bvals don't have the same number of entries.

If you have 34 volumes and 32 directions that implies that you have also acquired two b=0 images. You need to make sure that bvecs and bvals entries correspond to the same direction.


why? I removed the last gradient direction in bvecs and b value in bvals, respectively.


I saw many old  archives that we can use rotate_bvecs to rotate the bvecs after eddy_correct. Is it necessary? 
I didn't find such command in fsl ( I uses v4.10 version) . If rotate_bvecs is necessary, how should I do?

 
We no longer support the rotate_bvecs script, it is not clear whether rotating the bvecs or not makes any difference. In general, I would recommend going through the FDT practical from the FSL course, it will help you familiarize with the toolbox and the processing of DW images.
 
http://www.fmrib.ox.ac.uk/fslcourse/lectures/practicals/fdt/index.htm
 


Any comments,  Thank you in advance


Best


Shugao

 

 

Cheers,

Stam