mriconvert: Difference DWI Gradient Values

[Hardy, Peter A]

Fernando

I'm not sure I know where you got the gradients in table 1 but I've seen differences of the magnitude you show between tables 1 and 2 when we compare the prescribed gradient tables and the ultimate gradient values as listed in the header of the diffusion weighted images. I believe these differences arise from the inclusion of the imaging gradients into the calculation of the b-value. In our system this would happen at run time and would depend upon the specifics of the imaging acquisition such as FOV, slice thickness and slice orientation.

Peter Hardy

From: mriconvert-bounces at lists.uoregon.edu [mailto:mriconvert-bounces at lists.uoregon.edu] On Behalf Of Jolinda Smith
Sent: Wednesday, April 17, 2013 5:28 PM
To: mriconvert at lists.uoregon.edu; Fernando Garcia
Subject: Re: mriconvert: Difference DWI Gradient Values

Fernando,

Were your slices angled relative to the magnet coordinates? Generally, the "raw" directions are relative to the magnet coordinates, but most image processing software (such as FSL) requires the bvecs to be specified relative to the image coordinates. MRIConvert takes the gradients from the Dicom files and rotates them into the image frame, so unless you are using purely axial slices you expect them to be different.

It's a very, very good idea to double check this, especially any time something in your acquisitions-analysis chain changes (sequence, scanner software version, conversion software, etc). You can do this by taking a sample data set with oblique slices (the more oblique the better), running it through your analysis, and looking at the primary eigenvectors of the diffusion tensor (ie, look at an FA map color coded by the V1 directions). If the bvecs are wrong, it should be pretty obvious if the slice angles are oblique enough.

Jolinda

On Wed, 17 Apr 2013 06:23:19 -0700, Fernando Garcia <fernandogarcia.jm at gmail.com<mailto:fernandogarcia.jm at gmail.com>> wrote:


Dear MRIConverter Group,

I am converting DICOM data from a Siemens Espree 1.5T machine for Fiber Tracking. Looking at the converted files I see that there is a difference between the raw gradients and the gradients once converted. I am interested to know where these differences come from. What pre processing steps are performed on the data during conversion? Is it a correction? Is it recomended to do that instead of use the raw vector gradients?

I have included the gradient values I have obtained from the software here as reference. I have also summarized my findings below.

Your help in helping me to identify the source of these differences is much appreciated!!

Best Regards

Fernando
1. Raw Info of gradient

File: RawDicomGradient.txt (The gradients are not sort as the output of other tools)
0.000000 0.000000 0.000000
0.999977 0.006051 0.003027
-0.000006 -0.999982 0.006006
-0.024989 -0.653018 -0.756930
0.588672 0.769595 -0.247364
...

2.MRIConverter (dcm2nii)
File:MRIConverter.txt
dw_scheme:0.000000 0.000000 0.000000
dw_scheme:0.999660 -0.004982 -0.025588
dw_scheme:-0.008132 0.962882 -0.269799
dw_scheme:-0.052352 0.419028 -0.906463
dw_scheme:0.587622 -0.807974 -0.043334
...


--
Jolinda Smith, Ph.D.
MR Physicist
Interim Director of Operations
Lewis Center for Neuroimaging
University of Oregon
Eugene, OR USA
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists-prod.uoregon.edu/pipermail/mriconvert/attachments/20130418/9f7b74d9/attachment.html>