Was this ever answered?
1. Units - Not sure what the units of MD are. It might depend on your b-value 2. MD is not constrained to be between 0 and 1. Where are you getting that? 3. I don't know about the CC values. The CC ROIs are pretty small, and you might be getting significant partial voluming
doug
On 06/06/2014 02:06 AM, Rotem Saar wrote:
Dear freesurfer experts,
I'm using freesurfer 5.3 for extracting FA and MD values for each region from the *aseg file for each subject. While I get validated FA values (I validated them with you, the literature and some other software I used), I have a problem getting reasonable MD values.
This is the script I'm using for extracting the MD values:
- dt_recon --i
/usr/local/freesurfer/subjects/Anatomy0981/DTI/I160.dcm --s Anatomy0981 --o /usr/local/freesurfer/subjects/Anatomy0981/DTI --b /usr/local/freesurfer/subjects/Anatomy0981/DTI/bvals.bval /usr/local/freesurfer/subjects/Anatomy0981/DTI/bvecs.bvec
- mri_label2vol --seg
/usr/local/freesurfer/subjects/Anatomy0981/mri/wmparc.mgz --temp /usr/local/freesurfer/subjects/Anatomy0981/mri/wmparc.mgz --reg /usr/local/freesurfer/subjects/Anatomy0981/DTI/register.dat --o /usr/local/freesurfer/subjects/Anatomy0981/DTI/forMD.mgz
- mri_segstats --seg-erode 1 --seg
/usr/local/freesurfer/subjects/Anatomy0981/mri/wmparc.mgz --ctab $FREESURFER_HOME/FreeSurferColorLUT.txt --i /usr/local/freesurfer/subjects/Anatomy0981/DTI/forMD.mgz --sum /usr/local/freesurfer/subjects/Anatomy0981/stats/all_stats_MD-masked
It writes *.txt files similar to the ones of the FA (and also the volumetric analysis) which is good, but -
- I can't understand the units (it is written "unknown"). can u tell
me what do these units refer to ? 2) I can't understand why the units I get don't corresponds to the units in the literature (0-1). Is there a transformation I need to apply to these units in order to get the values we all see in the literature (MD values ranges between 0 to 1 in the papers I read). 3) I am concerned about the differences in the values both between and within subjects. For example, see 3 different subjects and their MD values for the corpus-callosum:
subject 1:
35 251 368 368.0 CC_Posterior 7.4783 7.9938 0.0000 16.0000 16.0000 36 252 110 110.0 CC_Mid_Posterior 56.7273 476.3758 0.0000 5002.0000 5002.0000 37 253 205 205.0 CC_Central 992.4537 1806.8883 0.0000 5002.0000 5002.0000 38 254 132 132.0 CC_Mid_Anterior 1664.2954 2328.8467 43.0000 5002.0000 4959.0000 39 255 277 277.0 CC_Anterior 1723.4620 1914.7883 0.0000 5002.0000 5002.0000
subject 2:
36 251 328 328.0 CC_Posterior 0.0000 0.0000 0.0000 0.0000 0.0000 37 252 78 78.0 CC_Mid_Posterior 43.9744 226.6371 0.0000 2000.0000 2000.0000 38 253 100 100.0 CC_Central 203.0800 779.8602 0.0000 5001.0000 5001.0000 39 254 83 83.0 CC_Mid_Anterior 264.5783 364.1581 0.0000 1002.0000 1002.0000 40 255 321 321.0 CC_Anterior 985.8879 720.5300 0.0000 3028.0000 3028.0000
subject 3:
35 251 424 424.0 CC_Posterior 20.6840 6.6236 10.0000 60.0000 50.0000 36 252 160 160.0 CC_Mid_Posterior 143.4312 780.4187 0.0000 5001.0000 5001.0000 37 253 202 202.0 CC_Central 538.0248 1143.9441 0.0000 5001.0000 5001.0000 38 254 179 179.0 CC_Mid_Anterior 496.5140 623.4088 0.0000 2023.0000 2023.0000 39 255 361 361.0 CC_Anterior 211.6094 539.0007 0.0000 2028.0000 2028.0000
the values are really different both within the CC parts of the same subject (does this make any sense ??) and also between subjects...
We ruled out a problem in the gradient matrix (both with you and because the FA values corresponds to reasonable values both from the literature and other software we use).
Can u please advise on how can we solve this issue ? We are in the middle of submission of a very novel study and were required to show addition analysis of the MD values - this is why this is so impotent to us. Attached please find the full *.txt files I got.
Thanks for your help.
Rotem Saar-Ashkenazy
Department of Brain and Cognitive Science Ben Gurion University of the Negev, Beer-Sheva, 84105
School of Social Work Ashkelon Academic College, Ashkelon, 78211
Israel
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer