Top | Previous

Tract Statistics

Remember...

For each new terminal that you open, you must do:

setenv SUBJECTS_DIR $TUTORIAL_DATA/diffusion_recons
cd $TUTORIAL_DATA/diffusion_tutorial

This section of the tutorial will teach you how to extract statistical measures for the major white-matter tracts reconstructed by TRACULA. There are two types of statistics files that are created for each white-matter tract:

• pathstats.overall.txt - This file contains measures averaged over the whole WM tract
• pathstats.byvoxel.txt - This files contains measures as a function of location along the length of the WM tract.

To view the statistics files from TRACULA for the left cortico-spinal tract of subject Diff001, cd into one subject's directory where the stats are located for this specific tract:

cd $TUTORIAL_DATA/diffusion_tutorial/Diff001/dpath/5cpts/priorcvs/seg14/initmni/regbbr/lh.cst_AS_avg23_cvs_bbr

pathstats.overall.txt

You can view the overall path stats by doing:

emacs $TUTORIAL_DATA/diffusion_tutorial/Diff001/dpath/5cpts/priorcvs/seg14/initmni/regbbr/lh.cst_AS_avg23_cvs_bbr/pathstats.overall.txt &

When looking at this file you will see something like this:

# Title Pathway Statistics
#
# generating_program $FREESURFER_HOME/bin/dmri_pathstats
# cvs_version 
# cmdline $FREESURFER_HOME/bin/dmri_pathstats --intrc $TUTORIAL_DATA/diffusion_tutorial/Diff001/dpath/5cpts/priorcvs/seg14/initmni/regbbr/lh.cst_AS_avg23_cvs_bbr --dtbase $TUTORIAL_DATA/diffusion_tutorial/Diff001/dmri/dtifit --path lh.cst --subj Diff001 --out $TUTORIAL_DATA/diffusion_tutorial/Diff001/dpath/5cpts/priorcvs/seg14/initmni/regbbr/lh.cst_AS_avg23_cvs_bbr/pathstats.overall.txt --outvox $TUTORIAL_DATA/diffusion_tutorial/Diff001/dpath/5cpts/priorcvs/seg14/initmni/regbbr/lh.cst_AS_avg23_cvs_bbr/pathstats.byvoxel.txt 
# sysname Linux
# hostname compute-0-1.local
# machine x86_64
# user rspriti
# anatomy_type pathway
#
# subjectname Diff001
# pathwayname lh.cst
#
Count 100
Volume 235
Len_Min 45
Len_Max 69
Len_Avg 55.3
Len_Center 53
AD_Avg 0.00108402
AD_Avg_Weight 0.00107521
AD_Avg_Center 0.00107062
RD_Avg 0.00046958
RD_Avg_Weight 0.000476065
RD_Avg_Center 0.000507009
MD_Avg 0.000674393
MD_Avg_Weight 0.000675781
MD_Avg_Center 0.000694881
FA_Avg 0.50122
FA_Avg_Weight 0.48976
FA_Avg_Center 0.456292

This text file contains various diffusion measures, averaged over the entire white-matter tract (for CST). The measures include:

• Number of sample paths in the WM tract
• Tract volume (in voxels)
• Maximum, minimum and average length of sample paths
• Length of the highest-probability (a.k.a. maximum a posteriori) path
• Axial diffusivity (average over the entire support of the path distribution, weighted average over the entire support of the path distribution, and average over highest-probability path only)
• Radial diffusivity (as above)
• Mean diffusivity (as above)
• Fractional anisotropy (as above)

pathstats.byvoxel.txt

Now take a look at the by voxel stats:

emacs $TUTORIAL_DATA/diffusion_tutorial/Diff001/dpath/5cpts/priorcvs/seg14/initmni/regbbr/lh.cst_AS_avg23_cvs_bbr/pathstats.byvoxel.txt &

It will look something like this:

# Title Pathway Statistics
#
# generating_program $FREESURFER_HOME/dmri_pathstats
# cvs_version 
# cmdline $FREESURFER_HOME/dmri_pathstats --intrc $TUTORIAL_DATA/diffusion_tutorial/Diff001/dpath/5cpts/priorcvs/seg14/initmni/regbbr/lh.cst_AS_avg23_cvs_bbr --dtbase $TUTORIAL_DATA/diffusion_tutorial/Diff001/dmri/dtifit --path lh.cst --subj Diff001 --out $TUTORIAL_DATA/diffusion_tutorial/Diff001/dpath/5cpts/priorcvs/seg14/initmni/regbbr/lh.cst_AS_avg23_cvs_bbr/pathstats.overall.txt --outvox $TUTORIAL_DATA/diffusion_tutorial/Diff001/dpath/5cpts/priorcvs/seg14/initmni/regbbr/lh.cst_AS_avg23_cvs_bbr/pathstats.byvoxel.txt 
# sysname Linux
# hostname compute-0-1.local
# machine x86_64
# user rspriti
# anatomy_type pathway
#
# subjectname Diff001
# pathwayname lh.cst
#
# pathway start
x y z AD RD MD FA
62 63 10 0.000637887 0.000476948 0.000530594 0.178865
62 63 11 0.0006074 0.000433505 0.00049147 0.22773
63 63 12 0.000846852 0.000610935 0.000689574 0.251636
63 64 13 0.00123916 0.00105259 0.00111478 0.107897
63 64 14 0.000936783 0.000621757 0.000726766 0.264886
64 64 15 0.000869413 0.000500918 0.00062375 0.348479
64 64 16 0.000742746 0.000432898 0.000536181 0.357518
65 64 16 0.000729103 0.000361411 0.000483975 0.42985
66 65 17 0.000645165 0.000105291 0.000285249 0.86012
67 65 18 0.00115478 0.000351219 0.000619072 0.647404
68 66 19 0.00142292 0.000690686 0.000934763 0.432432
69 66 20 0.00166259 0.000912869 0.00116278 0.363279
69 66 21 0.00145087 0.000448438 0.000782583 0.639517
70 66 22 0.00124617 0.000204701 0.000551856 0.816332
70 66 23 0.00138642 0.000415134 0.000738896 0.650657
71 66 23 0.00136252 0.000330065 0.000674217 0.721118
71 66 24 0.00134187 0.000385197 0.000704089 0.678151
71 66 25 0.00108983 0.000289931 0.000556565 0.729769
72 66 26 0.00106165 0.000357968 0.000592529 0.650427
72 66 27 0.00117242 0.000398032 0.00065616 0.619191
73 65 28 0.00130606 0.000441963 0.000729994 0.61662
73 65 29 0.00132239 0.00034928 0.000673651 0.701421
74 65 30 0.00133145 0.000391918 0.000705096 0.652609
74 65 31 0.00130265 0.000371743 0.000682045 0.665608
74 65 32 0.00126357 0.000429947 0.000707822 0.596254
75 65 33 0.00126146 0.000533254 0.000775991 0.499365
75 64 34 0.00117235 0.000465188 0.00070091 0.536046
75 64 35 0.00111144 0.000424831 0.000653701 0.560041
76 64 36 0.00128036 0.000299521 0.000626469 0.731721
76 64 37 0.00119495 0.000322181 0.000613105 0.688909
76 64 38 0.00105691 0.000334124 0.000575053 0.64994
76 63 39 0.00106784 0.000449252 0.000655449 0.514129
76 63 40 0.00113234 0.000537139 0.000735539 0.463226
76 63 41 0.000990485 0.000493054 0.000658864 0.446701
76 63 42 0.000798778 0.000477672 0.000584708 0.336071
76 62 43 0.000870652 0.000483175 0.000612334 0.362795
75 62 44 0.000878542 0.000493417 0.000621792 0.376378
75 62 45 0.000897465 0.000514559 0.000642194 0.348244
75 61 46 0.000882752 0.000537658 0.000652689 0.310019
74 61 47 0.00100985 0.00048457 0.000659663 0.436199
74 61 48 0.00104135 0.000461786 0.000654973 0.482846
74 60 49 0.000957188 0.000488566 0.000644774 0.472547
73 60 50 0.00100422 0.00055232 0.000702954 0.378698
73 60 51 0.000938077 0.000492242 0.000640854 0.429111
72 59 52 0.00104503 0.000475882 0.000665599 0.474443
72 59 53 0.00091483 0.00047261 0.000620017 0.409087
72 59 54 0.000792557 0.000632776 0.000686036 0.174565
71 59 55 0.000961412 0.000680667 0.000774249 0.243739
71 58 56 0.001061 0.000895478 0.000950651 0.133396
70 58 57 0.000862592 0.000663684 0.000729986 0.162365
70 58 58 0.00119199 0.00103056 0.00108437 0.086213
69 57 59 0.00120589 0.000980809 0.00105584 0.127098
69 57 60 0.00102806 0.000829161 0.000895461 0.141788
# pathway end

This text file contains various diffusion measures in each voxel along the highest-probability path. The first three entries in each row are the x, y, z coordinates of the voxel. The remaining entries are the axial diffusivity, radial diffusivity, mean diffusivity, and fractional anisotropy at that voxel.

path.overall.txt to Table Conversion

Measures can be extracted from these files to be analyzed further, e.g., for tract-based group analysis. Specifically, the text files can be converted into a table using the command tractstats2table for doing GLM analyses using mri_glmfit or any other statistical software (SPSS, Excel, Statview etc.)

To extract all diffusion measures into a table do the following:

tractstats2table --inputs $TUTORIAL_DATA/diffusion_tutorial/Diff001/dpath/5cpts/priorcvs/seg14/initmni/regbbr/lh.cst_AS_avg23_cvs_bbr/pathstats.overall.txt --overall --tablefile Diff001_LH.CST.All.Priorcvs.table

1. Notice we have specified the stats file with --inputs.
2. The flag, --overall expects your input to be path.overall.txt and converts all the measures specified in the text file into a table. Alternatively you can specify the --overall flag with --only-measures flag to extract specific measures from the tract and convert them to a table. (Another option is to use --byvoxel flag instead of the --overall flag
3. The output file will be called Diff001_LH.CST.All.Priorcvs.table as specified by the --tablefile flag.

Take a look at the resulting table file:

emacs $TUTORIAL_DATA/diffusion_tutorial/Diff001/dpath/5cpts/priorcvs/seg14/initmni/regbbr/lh.cst_AS_avg23_cvs_bbr/Diff001_LH.CST.All.Priorcvs.table

Instead of extracting all the measures from path.overall.txt for CST, we could grab just a few specific measures. For example, to extract only the Average Fractional Anisotropy along the Left Cortico-spinal tract, do the following:

tractstats2table --inputs $TUTORIAL_DATA/diffusion_tutorial/Diff001/dpath/5cpts/priorcvs/seg14/initmni/regbbr/lh.cst_AS_avg23_cvs_bbr/pathstats.overall.txt --overall --only-measures FA_Avg --tablefile Diff001_LH.CST.FA.Priorcvs.table

1. Notice the --only-measures flag followed by FA_Avg to indicate we are only interested in having the average FA over the path in table format for this path (CST).

For a group analysis that has multiple subjects, tractstats2table can be used to combine all the subjects' diffusion measures from the CST path.overall.txt files into a single table that can be used for analysis with mri_glmfit or any other statistics software

An easy way to do this is to create a text file that contains the names of the subjects and paths to their pathstats.overall.txt files. An example is given below:

$TUTORIAL_DATA/Diff001/dpath/priorcvs/regbbr/lh.cst_AS_avg23_cvs_bbr/pathstats.overall.txt
$TUTORIAL_DATA/Diff002/dpath/priorcvs/regbbr/lh.cst_AS_avg23_cvs_bbr/pathstats.overall.txt
$TUTORIAL_DATA/Diff003/dpath/priorcvs/regbbr/lh.cst_AS_avg23_cvs_bbr/pathstats.overall.txt
$TUTORIAL_DATA/Diff004/dpath/priorcvs/regbbr/lh.cst_AS_avg23_cvs_bbr/pathstats.overall.txt
$TUTORIAL_DATA/Diff005/dpath/priorcvs/regbbr/lh.cst_AS_avg23_cvs_bbr/pathstats.overall.txt
$TUTORIAL_DATA/Diff006/dpath/priorcvs/regbbr/lh.cst_AS_avg23_cvs_bbr/pathstats.overall.txt
$TUTORIAL_DATA/Diff007/dpath/priorcvs/regbbr/lh.cst_AS_avg23_cvs_bbr/pathstats.overall.txt
$TUTORIAL_DATA/Diff008/dpath/priorcvs/regbbr/lh.cst_AS_avg23_cvs_bbr/pathstats.overall.txt
$TUTORIAL_DATA/Diff009/dpath/priorcvs/regbbr/lh.cst_AS_avg23_cvs_bbr/pathstats.overall.txt
$TUTORIAL_DATA/Diff010/dpath/priorcvs/regbbr/lh.cst_AS_avg23_cvs_bbr/pathstats.overall.txt

Once this is created and saved as lh_cst_priorcvs_overall_pathstats, it can be passed as a parameter to the tractstats2table command. The following command will create a table with all the subjects specified in the lh_cst_priorcvs_overall_pathstats file and all the diffusion measures from pathstats.overall.txt for each subject.

tractstats2table --load-pathstats-from-file $TUTORIAL_DATA/diffusion_tutorial/TraculaStatFiles/lh_cst_priorcvs_overall_pathstats --overall --tablefile LH.CST.All.Priorcvs.table

To extract only specific measures like FA_Avg, do the following:

tractstats2table --load-pathstats-from-file $TUTORIAL_DATA/diffusion_tutorial/TraculaStatFiles/lh_cst_priorcvs_overall_pathstats --overall --only-measures FA_Avg --tablefile LH.CST.FA.Priorcvs.table

1. Note the use of the --only-measures flag again.

You can look at these tables in OpenOffice (or any other spreadsheet program). For example, to open the table LH.CST.All.Priorcvs.table in OpenOffice, do the following:

oocalc $TUTORIAL_DATA/diffusion_tutorial/TraculaStatFiles/LH.CST.All.Priorcvs.table 

From here, you could use these stats files to run a group analysis on the tracts.

Top | Previous