[Mrtrix-discussion] FiberCup Phantom Spherical Deconvolution

[Jeurissen Ben]

One of the nice things about the mrtrix pipeline is that it's not a black box solution, but that it allows you to inspect the outcome of each individual processing step using mrview.

Doing so, you will probably notice two things:

1) your "brain mask" is probably very poor as the brain mask example in the manual is designed with _brain_ data in mind and not a synthetic phantom

2) the fiber cup phantom has _no_ fibrous voxels with FA > 0.7, so your single fiber mask at best will be completely empty, which is exactly what estimate_response is telling you


________________________________
From: mrtrix-discussion-bounces at public.nitrc.org [mrtrix-discussion-bounces at public.nitrc.org] on behalf of Jesse Ross-Jones [jesse.rj at gmail.com]
Sent: 07 May 2013 11:55
To: mrtrix-discussion at public.nitrc.org
Subject: [Mrtrix-discussion] FiberCup Phantom Spherical Deconvolution

Dear MRtrix users/Devs,

I am trying to use the spherical deconvolution method in MRtrix to calculate the tracts in the FiberCup phantom available here: http://www.lnao.fr/spip.php?rubrique79. I am running the commands listed below, but am unable to successfully run estimate_response. Instead of showing the progress bar to 100% I simply get: estimate_response: estimating response function...  - ok

Do I need to modify some of the parameters or alter my approach in order to incorporate the separate gradients?

Many thanks for the help,
Jesse

the list of commands I am running are basically the default (except for using -grad to provide the modified gradient values):

mrconvert \FiberCup\6x6x6\dwi-b2650.nii dwi.mif

average dwi.mif -axis 3 - | threshold - - | median3D - - | median3D - mask.mif

dwi2tensor dwi.mif -grad diffusion_directions_corrected_2650.txt dt.mif

tensor2FA dt.mif - | mrmult - mask.mif fa.mif

tensor2vector dt.mif - | mrmult - fa.mif ev.mif

erode mask.mif -npass 3 - | mrmult fa.mif - - | threshold - -abs 0.7 sf.mif

estimate_response dwi.mif sf.mif -grad diffusion_directions_corrected_2650.txt response.txt

csdeconv dwi.mif response.txt -lmax 8 -mask mask.mif CSD8.mif

streamtrack SD_PROB CSD8.mif -seed seed.mif -mask mask.mif cst_csd.tck

A sample from diffusion_directions_corrected_2650.txt:

0       0       0       0
1       0       0       2650
0       -0.987414       -0.158158       2650
-0.026007       -0.761231       0.64796 2650

The problem I run into is running estimate_response (following is the output -debug):
estimate_response dwi.mif sf.mif -grad diffusion_directions_corrected_2650.txt response.txt -debug
estimate_response [DEBUG]: reading key/value file "C:\mrtrix.conf"...
estimate_response [INFO]: opening image "dwi.mif"...
estimate_response [DEBUG]: reading key/value file ".\dwi.mif"...
estimate_response [DEBUG]: sanitising transformation matrix...
estimate_response [DEBUG]: preparing file ".\dwi.mif"
estimate_response [DEBUG]: setting up image "dwi.mif"...
estimate_response [DEBUG]: sanitising transformation matrix...
estimate_response [DEBUG]: setting up data increments for "dwi.mif"...
estimate_response [DEBUG]: data increments initialised with start = 0, stride =
[ 1 64 -4096 4096 ]
estimate_response [INFO]: opening image "sf.mif"...
estimate_response [DEBUG]: reading key/value file ".\sf.mif"...
estimate_response [DEBUG]: sanitising transformation matrix...
estimate_response [DEBUG]: preparing file ".\sf.mif"
estimate_response [DEBUG]: setting up image "sf.mif"...
estimate_response [DEBUG]: sanitising transformation matrix...
estimate_response [DEBUG]: setting up data increments for "sf.mif"...
estimate_response [DEBUG]: data increments initialised with start = 0, stride =
[ 1 64 -4096 ]
estimate_response [INFO]: found 65x4 diffusion-weighted encoding
estimate_response [INFO]: found b=0 images in studies [ 0 ]
estimate_response [INFO]: found 64 diffusion-weighted directions
estimate_response [INFO]: calculating even spherical harmonic components up to o
rder 8
estimate_response [DEBUG]: mapping image "dwi.mif"...
estimate_response [DEBUG]: file ".\dwi.mif" mapped at 0x21e0000, size 532750 (re
ad-only)
estimate_response [DEBUG]: data mapper for image "dwi.mif" mapped with segment s
ize = 266240
estimate_response [DEBUG]: mapping image "sf.mif"...
estimate_response [DEBUG]: file ".\sf.mif" mapped at 0x3d0000, size 768 (read-on
ly)
estimate_response [DEBUG]: data mapper for image "sf.mif" mapped with segment si
ze = 4096
estimate_response [INFO]: found 0 single-fibre voxels
estimate_response: estimating response function...  - ok
estimate_response [INFO]: estimated response SH coefficients: [ nan nan nan nan
nan ]
estimate_response [INFO]: closing image "dwi.mif"...
estimate_response [DEBUG]: unmapping file ".\dwi.mif"
estimate_response [INFO]: closing image "sf.mif"...
estimate_response [DEBUG]: unmapping file ".\sf.mif"
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.nitrc.org/pipermail/mrtrix-discussion/attachments/20130507/00a10d08/attachment.html