According to the bedpostx_datacheck info you posted earlier, your resolution is 0.5x0.5x2.4.  This is also not ideal for tractography as your ability to resolve the superior to inferior component of fibers in much worse than your ability to resolve the other components.  Better is 2mmX2mmX2mm, which has both good resolution and SNR.  At a minimum you should turn off interpolation on the scanner (and then your resolution would be 1x1x2.4mm).

 

Peace,

 

Matt.

 

From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf Of Mi Jung Kim
Sent: Wednesday, March 11, 2009 11:02 AM
To: [log in to unmask]
Subject: Re: [FSL] capillary phantom for muscle dti2

 

Hello,

When you run FDT, how isotropic voxel has to be? I am dealing with 1x1x2.4. Is it too non-isotropic?

Thanks,

Mi Jung

On Wed, Mar 11, 2009 at 8:52 AM, Mi Jung Kim <[log in to unmask]> wrote:

I got you.

But my SNR is pretty high. How high SNR is required when I have few gradient directions for this algorithm. I know that it is recommended use more than 25 gradient directions for this calculation. Do you think that getting more gradient directions is only way to solve this isse?

Thanks,

Mi Jung

 

On Wed, Mar 11, 2009 at 8:45 AM, Saad Jbabdi <[log in to unmask]> wrote:

DTIFIT does not impose that the fitted tensor is positive definite, so you may have negative eigenvalues (i.e. negative Ls, which can produce FA>1). 

Since you only have 6 data points to calculate your tensor, your chances of getting negative tensors are higher, esp if SNR is low..

 

 

Saad.

 

 

On 11 Mar 2009, at 15:40, Mi Jung Kim wrote:



Yes, I can see V1 and FA. But when I check the values with fslstats they seemed to be odd to me. Because FA has max of 1, but it was more than 1 something like that. Here is my question. I do not think that lamdas can have negative values, but fslstats shows them in negative min as well. Why is that?

Thanks,

Mi Jng

On Wed, Mar 11, 2009 at 8:22 AM, Saad Jbabdi <[log in to unmask]> wrote:

They look all right to me.

Why are you saying they are not right? 

Did you have a look at FA and V1 in fslview?

 

Saad.

 

On 11 Mar 2009, at 15:18, Mi Jung Kim wrote:



Hi,

These are my outputs from dtifit.
I did a carriage return in bvals.
Are they right? They do not seem to be right values to me.
Any thought?

Sincerely,

Mi Jung
computer-four:~/data macadmin$ fslstats dti_FA.nii.gz -R
0.000000 1.224744
computer-four:~/data macadmin$ fslstats dti_V1.nii.gz -R
-0.999986 0.999999
computer-four:~/data macadmin$ fslstats dti_L1.nii.gz -R
-0.000893 0.003149
computer-four:~/data macadmin$ fslstats dti_L2.nii.gz -R
-0.001136 0.002519
computer-four:~/data macadmin$ fslstats dti_L3.nii.gz -R
-0.002149 0.002262
computer-four:~/data macadmin$ fslstats dti_V2.nii.gz -R
-0.999982 0.999999
computer-four:~/data macadmin$ fslstats dti_V3.nii.gz -R
-0.999986 1.000000
computer-four:~/data macadmin$ fslstats dti_MD.nii.gz -R
-0.001113 0.002447
computer-four:~/data macadmin$ fslstats dti_S0.nii.gz -R
0.000000 16406.000000

On Wed, Mar 11, 2009 at 8:01 AM, Mark Jenkinson <[log in to unmask]> wrote:

Hi,

Just add a blank line at the end with a text editor.

All the best,
       Mark




On 11 Mar 2009, at 14:49, Mi Jung Kim wrote:

Hi Sadd,

How can I add a carriage return?

Thanks,

Mi Jung

On Wed, Mar 11, 2009 at 1:54 AM, Saad Jbabdi <[log in to unmask]> wrote:
Yes, it sounds like you have a version which still had the bvals bug.
Just add a carriage return at the end of your bvals file and that should solve the problem.

Saad.

On 10 Mar 2009, at 18:00, Mi Jung Kim wrote:

Hi,

I am using FSL4.0 on Mac.

Thanks,

Mi Jung

On Tue, Mar 10, 2009 at 10:19 AM, Saad Jbabdi <[log in to unmask]> wrote:
I suspect this is because you need to add a carriage return at the end of your bvals file (which version of FSL are you using?)

Saad.


On 10 Mar 2009, at 16:29, Mi Jung Kim wrote:

Hi,

I have been working capillary phantom for muscle DTI.

I just was able to run all procedures, but after dtifit, outputs are
blank..Do you have any thought on that?

One more question right now is that since I am dealing with phantom data, do
I still go though all steps before probtrack?


Thanks,

Mi Jung


Saad Jbabdi
Oxford University FMRIB Centre

JR Hospital, Headington, OX3 9DU, UK
+44 (0) 1865 222545  (fax 717)
www.fmrib.ox.ac.uk/~saad


Saad Jbabdi
Oxford University FMRIB Centre

JR Hospital, Headington, OX3 9DU, UK
+44 (0) 1865 222545  (fax 717)
www.fmrib.ox.ac.uk/~saad






 

 

Saad Jbabdi

Oxford University FMRIB Centre

 

JR Hospital, Headington, OX3 9DU, UK

+44 (0) 1865 222545  (fax 717)

 

 

 

 

 

 

 

Saad Jbabdi

Oxford University FMRIB Centre

 

JR Hospital, Headington, OX3 9DU, UK

+44 (0) 1865 222545  (fax 717)