You must be logged in to post Login Register


Lost Your Password?

Search Forums:


 






Wildcard Usage:
*    matches any number of characters
%    matches exactly one character

Tensor reconstruction with GE DWI data

UserPost

10:52 am
March 11, 2010


cibuthomas

Member

posts 16

Hello

I have DWI data in 60 direction that were collected in 4 runs. DTK recognizes the data as 4 unique series. Is there anyway I can use DTK to combine the data into a single series so that DTK can do the reconstruction?

Thank you

Cibu

1:58 pm
March 11, 2010


Ruopeng

Admin

posts 406

Post edited 6:59 pm – March 11, 2010 by Ruopeng


Hi Cibu,

In that case, you will have to convert the images into individual 3D volumes and name them with numbered suffix, such xxx001.nii, xxx002.nii… xxx060.nii… DTK can take a series of 3D volume files.

Ruopeng

4:33 pm
March 12, 2010


cibuthomas

Member

posts 16

Thanks Roupeng

Trouble is the b = 0 volumes are scattered across the 4 runs. Anyway, I used TORTOISE to get a single analyze volume with the b = 0 volume (N=10) first followed by b = 300 (N = 10) and b max = 1100 (N = 60 directions). I have the raw data set in FSL format, so I have the bvecs and bval files in seperate files. My question is how can I make a gradient table using the bvals and bvecs so that DTK understands the raw data.

 

cheers

Cibu

10:57 pm
March 14, 2010


Ruopeng

Admin

posts 406

The current version of DTK can not process data set with multiple high b values. Also, it requires all the b0 volumes at the top of the series. But the upcoming release (coming very soon!) is going to support multiple high b values as well as b0 volumes at anywhere in the series.

Ruopeng

5:03 pm
March 16, 2010


Ruopeng

Admin

posts 406

You might have noticed that a new version of Diffusion Toolkit has been released. This version will allow you to put b0 vector in the gradient table. You can also specify b value as the 4th component in the table. You still however need to number your image files correctly in this case because they are not from one series.

Ruopeng

10:35 pm
October 19, 2010


pguillod

New Member

posts 2

Post edited 2:40 am – October 20, 2010 by pguillod


I'm sure someone already did this, but if you're still trying to convert bvals and bvecs from the diffusion toolbox in FSL to the format used for DTK just write the following lines into a script (i.e. bvecs2grad), add it to your bin folder, and then when you're in the folder containing bvals and bvecs execute something per below. Then when you're in the DTK GUI under the gradient file click import, select the file, and name the table accordingly.

:: bvecs2grad mygradfile.something

#!/bin/sh
if [ $# = 1 ] ; then out=$1 ; else out=default.dat ; fi
if [ -f $out ] ; then rm $out ; fi #remove old file
x=(`sed -n '1p' bvecs`)
y=(`sed -n '2p' bvecs`)
z=(`sed -n '3p' bvecs`)
i=0
for bv in $(cat bvals) ; do
echo ${x[$i]}, ${y[$i]}, ${z[$i]}, $bv >> $out
i=`echo $i+1|bc`
done

Also, if you have multiple acquisitions for one set you can merge the files into a single image set using the fslmerge command. If my images are part1.nii and part2.nii then you can merge them by typing

fslmerge -t allofthem part1 part2