Very simply, if you had such a [[http://dti-tk.sourceforge.net/pmwiki/uploads/Documentation/dti_volume_list.txt|dti_volume_list.txt]] and wanted to perform group comparison without any covariation, the design matrix should be as followed ([[http://dti-tk.sourceforge.net/pmwiki/uploads/Documentation/design.txt|design.txt]]) and the contrast matrix should be as followed ([[http://dti-tk.sourceforge.net/pmwiki/uploads/Documentation/contrast.txt|contrast.txt]]).
 
 
 

Very simply, if you had such a [[http://dti-tk.sourceforge.net/pmwiki/uploads/Documentation/subjs_diffeo.txt|subjs_diffeo.txt]] and wanted to perform group comparison without any covariation, the design matrix should be as followed ([[http://dti-tk.sourceforge.net/pmwiki/uploads/Documentation/design.txt|design.txt]]) and the contrast matrix should be as followed ([[http://dti-tk.sourceforge.net/pmwiki/uploads/Documentation/contrast.txt|contrast.txt]]).
 
 
 

Very simply, if you had such a [[http://dti-tk.sourceforge.net/pmwiki/uploads/Documentation/subjs_diffeo.txt|dti_volume_list.txt]] and wanted to perform group comparison without any covariation, the design matrix should be as followed ([[http://dti-tk.sourceforge.net/pmwiki/uploads/Documentation/design.txt|design.txt]]) and the contrast matrix should be as followed ([[http://dti-tk.sourceforge.net/pmwiki/uploads/Documentation/contrast.txt|contrast.txt]]).
 
 
 

For this, use the script called [[http://dti-tk.sourceforge.net/pmwiki/uploads/Documentation/tsa_stats_glm|tsa_stats_glm]]. For this script to work, you must first generate contrast and design matrices to identify which subjects belong to which group.  
You can find a simple explanation and example of the general linear model, as well as the associated design and contrast matrix on the [[http://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/GroupAnalysis#DesignMatrix.2BAC8-FSGDFile|freesurfer website]]. 
Very simply, if you had such a [[http://dti-tk.sourceforge.net/pmwiki/uploads/Documentation/subjs_diffeo.txt|dti_volume_list.txt]] and wanted to perform group comparison without any covariation, the design matrix should be as followed ([[design.txt|http://dti-tk.sourceforge.net/pmwiki/uploads/Documentation/design.txt) and the contrast matrix should be as followed ([[contrast.txt|http://dti-tk.sourceforge.net/pmwiki/uploads/Documentation/contrast.txt]]).
 
 
 
An applet to automate this process can be found [[Documentation.matrix_generator|here]]. Once this is done, simply run the script in the given directory. Example:
 

The script ''tsa_sampling'', located in the dtitk/scripts directory of the DTI-TK package samples white matter and tract-specific attributes across the whole population. These white matter attributes are values, such as FA, RD, AD, ADC.... For each one of these values, the mean or the max are computed locally and projected onto the medial model of the tracts. A medial model or mesh is a sheet-like structure that represents a 3D tract (there are 11 of them in our atlases). More precisely, for each subject, these values are sampled along the spokes corresponding to each vertex of the mesh and projected on the mesh.

* maxFA or mean: the values projected on the mesh can be computed in two ways. The resulting projected values can be 1/ an average of similar values along the spokes of the vertex (mean) or 2/a value equal to the maximum value existing along these spokes (note: like mean, and unlike its name, the option maxFA will compute and project the max of all the WM values -FA, AD, RD, ADC- on the tract, not only the max of FA)
 

* feature: is the white matter attribute analyzed. It can be FA, Radial diffusivity (RD), apparent diffusion coefficient (ADC), axial diffusivity (AD) 
* p threshold is typically 0.01 or 0.05: 
* number of permutation: this number can be set to 0 to test the script, and raised to typically 10000

Non-parametric permutation-based supratheshold statistical analysis follows. In this last computation, we  use a script from dti-tk that completely relies on the function ''mesh_glm'', which is part of [[http://www.itksnap.org/pmwiki/pmwiki.php?n=CMREP.Documentation|cmrep]].

* tsa_model_path is the directory that contains all the medial representations (if you downloaded our template, this path should be something like "${HOME_DIR}/template/aging/tsa/"
* maxFA or mean: the value projected on the mesh can be computed in two ways. The resulting projected value can be 1/ an average of similar values along the spokes of the vertex (mean) or 2/a value equal  to the maximum FA existing along these spokes
 

Non-parametric permutation-based supratheshold statistical analysis follows. In this last computation, we  use a script from dti-tk that completely relies on the function "mesh_glm", which is part of [[http://www.itksnap.org/pmwiki/pmwiki.php?n=CMREP.Documentation|cmrep]].

* tsa_model_path is the directory that contains all the medial representations (if you downloaded our template, this path should be something like ''${HOME_DIR}/template/aging/tsa/"

The script ''tsa_sampling'', located in the dtitk/scripts directory of the DTI-TK package samples white matter values across the whole population. These values are scalar features, such as FA or MD. The average or the maximum of these values are computed locally and projected onto the medial model of the tracts. A medial model or mesh is a sheet-like structure that represents a 3D tract (there are 11 of them in our atlases). More precisely, for each subject, these values are sampled along the spokes corresponding to each vertex of the mesh and projected on the mesh.

The script ''tsa_sampling'', located in the dtitk/scripts directory of the DTI-TK package samples white matter values across the whole population. These values are scalar features, such as FA or MD. The average or the maximum of these values are computed locally and projected onto the medial model of the tracts. A medial model or mesh is a sheet-like structure that represents a 3D tract (there are 11 of them in our atlases). More precisely, for each subject, these values are sampled along the spokes corresponding to each vertex and projected on the mesh.

#[[http://www.itksnap.org/pmwiki/pmwiki.php?n=CMREP.Documentation|cmrep]], which is needed for the  statistical analysis

#[[http://www.itksnap.org/pmwiki/pmwiki.php?n=CMREP.Documentation|cmrep]], which is needed for the last step '''Statistical Analysis''''

#[[Downloads.Downloads|DTI-TK]]
#[[http://www.itksnap.org/pmwiki/pmwiki.php?n=CMREP.Documentation|cmrep]], which is needed for the last step

For this, use the script called [[http://dti-tk.sourceforge.net/pmwiki/uploads/Documentation/tsa_stats_glm|tsa_stats_glm]]. For this script to work, you must first generate contrast and design matrices to identify which subjects belong to which group. An applet to automate this process can be found [[Documentation.matrix_generatorl|here]]. Once this is done, simply run the script in the given directory. Example:
 

For this, use the script called [[http://dti-tk.sourceforge.net/pmwiki/uploads/Documentation/tsa_stats_glm|tsa_stats_glm]]. For this script to work, you must first generate contrast and design matrices to identify which subjects belong to which group. A script to automate this process can be found [[http://dti-tk.sourceforge.net/pmwiki/uploads/testing/config_file_generator.html|here]]. Once this is done, simply run the script in the given directory. Example:
 

* feature: is the white matter attribute analyzed. It can be FA, Radial diffusivity (RD), apparent diffusion coefficient (ADC), axial diffusivity (AD)

* mesh_in.txt is a text file containing the names of all the tracts (an example can be found [[http://dti-tk.sourceforge.net/pmwiki/uploads/Documentation/mesh_in.txt|here]])

For this, use the script called [[http://dti-tk.sourceforge.net/pmwiki/uploads/Documentation/tsa_stats_glm|tsa_stats_glm]]. For this script to work, you must first generate contrast and design matrices to identify which subjects belong to which group. A script to automate this process can be found [[http://dti-tk.sourceforge.net/pmwiki/uploads/Applet/config_file_generator.html|here]]. Once this is done, simply run the script in the given directory. Example:
 
->[@
tsa_stats_glm mesh_in.txt design.txt contrast.txt [feature] [p threshold] [number of permutation]
@]
 
where: 
* mesh_in.txt is in [[[[http://dti-tk.sourceforge.net/pmwiki/uploads/Documentation/mesh_in.txt|this]] format
* feature: is the white matter attribute analyzed. It can be FA, Radial diffusivity (RD), ADC, AD

* maxFA or mean: the value projected on the mesh can be computed in two ways. The resulting projected value can be 1/ an average of similar values along the spokes of the vertex (mean) or 2/a single value corresponding to the maximum FA.
 

* dti_volume_list.txt is a text file that gathers all the subjects'names (click [[http://dti-tk.sourceforge.net/pmwiki/uploads/Documentation/subjs_diffeo.txt|here]] for an example)

!!Sampling values over all the subjects
The script ''tsa_sampling'', located in the dtitk/scripts directory of the DTI-TK package samples white matter value across the whole population. A medial model or mesh is a sheet-like structure that represents a 3D tract. It is thus essential to project to compute the diffusion tensor value (or any other white matter attribute) at each point of this tract. To do so, for each subject, these values are sampled along the spokes corresponding to each vertex and projected on the mesh.
 
You can run it as follows:
 
->[@
tsa_sampling dti_volume_list.txt tsa_model_path [maxFA or mean]
@]
 
 
where
* dti_volume_list.txt is a text file that gathers all the subjects'names (click [[http://dti-tk.sourceforge.net/pmwiki/uploads/Documentation|here]] for an example)
* tsa_model_path is the directory that contains all the medial representations
* maxFA or mean: the value projected on the mesh can be computed in two ways: the resulting projected value can be an average of the 3D structure or a single value corresponding to the maximum FA>
 

!!Statistical Analysis
Non-parametric permutation-based supratheshold statistical analysis follows.
For this, use the script called [[http://dti-tk.sourceforge.net/pmwiki/uploads/Applet/group_tsa_stats.sh|group_tsa_stats]] (which needs to be modified depending on the threshold you want to test for, as well as your input and output meshes). For this script to work, you must first generate contrast and design matrices to identify which members belong to which group. A script to automate this process can be found [[http://dti-tk.sourceforge.net/pmwiki/uploads/Applet/group_tsa_stats.sh|here]]. Once this is done, simply run the group_tsa_stats script in the given directory.

->4. create final normalized volumes for analysis
->[@
dti_warp_to_template_group subjs.txt template.nii.gz 2 2 2
@]
 
->This step combines the affine and nonlinear transformations into a single transformation to warp the original data into the template space.  It allows you to specify the voxel spacing of the normalized data.  The example given sets them to 2x2x2 mm^3.
 

->subjs.txt is a text file containing the names of all the image files (see [[Attach:subjs.txt|here]]).
 
->''Output'': For each subject, subject_name.aff, a text file that contains an initialization matrix for the next step.
 

->keep the same subjs.txt. 
 
->''Output'': subject_name_aff.nii.gz
 
->3. nonlinear registration: 
->[@
dti_diffeomorphic_sn template.nii.gz subjs_aff.txt mask.nii.gz 6 0.002
@]
 
->subjs_aff.txt consists of a list of all the affinely registered subjects' names.
 
->''Output'': subject_name_aff_diffeo.nii.gz
 
 

Given the list in subjs_diffeo.txt, you can determined where the 0 and the 1 have to go in surfcluster.maxFA.config.txt.

This is the point, where we are going to use DTI-TK to register all the images to the common template. For this, you can refer to this [[Documentation.Registration|tutorial]], which walks you through the successive types registration: rigid, affine and nonlinear. It also gives you three distinct commands depending on if you want to build a template while registering your data. In our case, we already have the template, so the relevant commands to use are
 

subjs.txt is a text file containing the names of all the image files (see [[Attach:subjs.txt|here]]).
 
''Output'': For each subject, subject_name.aff, a text file that contains an initialization matrix for the next step.
 

dti_affine_sn template.nii.gz subjs.txt EDS 1

*an adult template, built with 78 subjects, 40 males/38 females (mean age: 39.5 ± 12), not yet published
* an aging template, generated from 51 subjects, 21 males/30 females (mean age: 70.1 ± 4), see Zhang et al., 2010, [[Publications.Publications|here]]
 

 
After going through this section, all your subjects will be in a  common space. Here again, there's no harm in checking the registered images to make sure everything is ok!
 

Non-parametric permutation-based suprathreshold statistical analysis follows. 

In this section, you will use the script tsa_sampling.sh, located in the dtitk/scripts directory of your DTI-TK package. It is a pretty straightforward shell scripts, which you will need to modify. 
 
# At line 22, the path tsa_model_path needs to be edited and must point to the location of the medial representations.
 
# That being done, create a .txt file containing the names of the registered subjects' files, such as dti_volume_list.txt
 
# One this is done, just run the command: tsa_sampling dti_volume_list.txt.
 
 
''Note'': If you do not work with a cluster or your system does not allow to `qsub' jobs, then line 25 of tsa_sampling.sh shall be changed as well.

and do so for each subject separately. After going through this section, all your subjects will be in a  common space. Here again, there's no harm in checking the registered images to make sure everything is ok!
 

 
->1. rigid registration: 

->2. affine registration:

->3. nonlinear registration: 

1. rigid registration: 

2. affine registration:

3. nonlinear registration: 

dti_affine_sn template.nii.gz subjs.txt EDS

 

subjs_aff.txt consists of a list of all the affinely registered subjects'names).  
 
''Output'': subject_name_aff_diffeo.nii.gz
 

subjs.txt is a text file containing the names of all the image files.
''Output'': subject_name.aff, a text file that contains an initialization matrix for the next step. 
 

 
keep the same subjs.txt. 
''Output'': subject_name_aff.nii.gz
 

This is the point, where we are going to use DTI-TK to register all the images to the common template. For this, you can refer to this [[Documentation.Registration|tutorial]], which walks you through the successive types registration: rigid, affine and nonlinear. It also gives you three distinct commands depending on if you want to build a template while registering your data. In our case, we already have the template, so the main commands to consider are

dti_rigid_reg template.nii.gz tensor.nii.gz EDS 4 4 4 0.01
dti_affine_reg template.nii.gz tensor.nii.gz EDS 4 4 4 0.01
dti_diffeomorphic_reg template.nii.gz tensor_aff.nii.gz mask.nii.gz 1 6 0.002

In this section, you will use the script tsa_sampling.sh, located in the dtitk/scripts of your DTI-TK package.

Here are the DTI and TSA representation of the atlas.
%center%%width=750px%Attach:atlas_example1.png

# Then you want to make sure that the diffusivity units of the DT images are compatible with DTI-TK. Check [[Documentation.Diffusivity|this page]] to make sure

1. Obtain an existing DTI template (they are not freely available for download yet, but will be soon. In the meantime, please send an email to [[mailto:cbrun@picsl.upenn.edu|cbrun@picsl.upenn.edu]]). Each template consists of a diffusion tensor atlas as well as medial representations of 11 tracts: the corpus callosum (CC), the corticospinal tracts (CST), inferior fronto-occipital tracts (IFO), inferior longitudinal tracts (ILF), superior longitudinal tracts (SLF), and uncinates (UNC). 
 

2. Register all the population diffusion tensor images to this template (rigid, affine and nonlinear registration)
3. Sample the Fractional Anisotropy (or Apparent Diffusion Coefficient values) over all the registered subjects, compute the mean of this value at each point and project this value on the medial representation
4. Perform a permutation-based non-parametric suprathreshold cluster analysis, to determine the WM clusters that are significantly different between your two populations.
 
 

This abstract explains the core of the method and reflects upon another paper, by Yushkevich et al. which was published two years earlier (see [[Publications.Publications|here]]). In this paper, the author built a children template.  From the two papers, you can easily guess the pipeline of the TSA technique.
Here is in a simplified version:
#Obtain an existing DTI template (they are not freely available for download yet, but will be soon. In the meantime, please send an email to [[cbrun@picsl.upenn.edu|cbrun@picsl.upenn.edu]]). Each template consists of a diffusion tensor atlas as well as medial representations of 11 tracts: the corpus callosum (CC), the corticospinal tracts (CST), inferior fronto-occipital tracts (IFO), inferior longitudinal tracts (ILF), superior longitudinal tracts (SLF), and uncinates (UNC).  
 
Two templates have been built from the IXI dataset that is freely available:

#[[http://www.paraview.org|Paraview]], a visualization tool
 
 

!!Computing Mean FA or Mean ADC

# rigid registration:  
->[@
dti_rigid_sn template.nii.gz subjs.txt EDS 
@]
 
(subjs.txt is a text file containing the names of all the image files). Output: subject\_name.aff, a text file that contains an initialization matrix for the next step. 
->[@
affine registration: dti_affine_sn template.nii.gz subjs.txt EDS
 @]
(keep the same subjs.txt). 

 
# nonlinear registration:  
->[@
dti_diffeomorphic_sn template.nii.gz subjs_aff.txt mask.nii.gz 6 0.002
@]
this time
subjs_aff.txt consists of a list of all the affinely registered subjects'names). Output: subject_name_aff_diffeo.nii.gz
 

@]
 

This abstract explains the core of the method and reflects upon another paper, by Yushkevich et al. which was published two years earlier (see [[Publications.Publications|here]]). In this paper, the author built a children template.  From the two papers, you can easily guess the pipeline of the TSA technique. Here is in a simplified version:
#Obtain an existing DTI template (they are not freely available for download yet, but will be soon. In the meantime, please send an email to [[cbrun@picsl.upenn.edu|cbrun@picsl.upenn.edu]]). Each template consists of a diffusion tensor atlas as well as medial representations of 11 tracts: the corpus callosum (CC), the corticospinal tracts (CST), inferior fronto-occipital tracts (IFO), inferior longitudinal tracts (ILF), superior longitudinal tracts (SLF), and uncinates (UNC). Two templates have been built from the IXI dataset that is freely available:

* an aging template, generated from 51 subjects, 21 males/30 females (mean age: 70.1 pm 4), see Zhang et al., 2010, [[Publications.Publications|here]]
 

%center% ''Left'': Principal diffusion direction map of the DTI template visualized with [[http://www.itksnap.org/|itk-snap]]. Blue represents the fibers that are going from inferior to superior regions, green posterior anterior and red from left to right. ''Right'': Medial representation of the 11 tracts: the cc (red), the CSTs (blue), the IFOs (yellow), the ILFs (green) the SLFs (purple) and UNCs (light blue, visualized with [[http://www.paraview.org/paraview]]).
 

This abstract explains the core of the method and reflects upon another paper, by Yushkevich et al. which was published two years earlier (see [[Publications.Publications|here]]. In this paper, the author built a children template.  From the two papers, you can easily guess the pipeline of the TSA technique. Here it is in a simplified version:

Here is the abstract from the very good paper published by Zhang et al. in the Proceedings of the 4'^th^' International Conference on Biomedical Image registration in 2010 (see the paper [[Publications.Publications|here]])
 

!!Prior requirement: Checking your DTIs 

Before starting this analysis, make sure you have these two softwares available.  
 
#[[Downloads.Downloads|DTI-TK]] 

Before starting this analysis, make sure you have these two softwares available. #[[Downloads.Downloads|DTI-TK]] 

%center%%width=700px%Attach:atlas_example1.png 
%center% ''Left'': Principal diffusion direction map of the DTI template. Blue represents the fibers that are going from inferior to superior regions, green posterior anterior and red from left to right. ''Right'': Medial representation of the 11 tracts: the cc (red), the CSTs (blue), the IFOs (yellow), the ILFs (green) the SLFs (purple) and UNCs (light blue).
 

%center%%width=600px%Attach:atlas_example1.png
 

%width=300px%Attach:atlas_example1.png
 

%center width=300px%Attach:atlas_example1.png
 

%width=40px%Attach:atlas_example1.png
 

%width= 40px%Attach:atlas_example1.png
 

%center width= 40px%Attach:atlas_example1.png

%center width: 20px%Attach:atlas_example1.png

%center width: 30px%Attach:atlas_example1.png

%center width: 50px%Attach:atlas_example1.png

%center width:16cm%Attach:atlas_example1.png

%center%Attach:atlas_example1.png

#affine registration: dti_affine_sn template.nii.gz subjs.txt EDS (keep the same subjs.txt).  
Output: subject_name_aff.nii.gz
# nonlinear registration: dti_diffeomorphic_sn template.nii.gz subjs_aff.txt mask.nii.gz 6 0.002 (this time subjs_aff.txt consists of a list of all the affinely registered subjects'names). Output: subject_name_aff_diffeo.nii.gz
 

Before starting this analysis, make sure you have these two softwares available. #[[Downloads.Downloads| DTI-TK]]  
#[[http://www.paraview.org|Paraview]], a visualization tool, which you can obtain from [[http://www.paraview.org/|here]]
 
 

# Obtain an existing DTI template (they are not freely available for download yet, but will be soon. In the meantime, please send an email to cbrun@picsl.upenn.edu). Each template consists of a diffusion tensor atlas as well as medial representations of 11 tracts: the corpus callosum (CC), the corticospinal tracts (CST), inferior fronto-occipital tracts (IFO), inferior longitudinal tracts (ILF), superior longitudinal tracts (SLF), and uncinates (UNC).

Before any statistical analysis, or registration, you must have built your Diffusion Tensor images. For them to work seamlessly with DTI-TK, there are a few things you might want to check.
# How were my images reconstructed? [[Documentation.Format|Here]] are a few tools that exist amongst others, which allow you to build your DTI from your DWI.
# Then you want to make sure that the diffusivity units of the DT images are compatible with DTI-TK. Check [[Documentation.Diffusivity|this page}} to make sure
# Last but not least, it is always a good idea (well, it's in fact highly recommended!) to visualize your images to make sure there is no outlier, which could critically influence your final analysis in a negative way. You will find a few very useful [[Documentation.VisualizationTool|commands]] in this toolkit that will enable you to check these images.
 

%center% ''Left'': Principal diffusion direction map of the DTI template. Blue represents the fibers that are going from inferior to superior regions, green posterior anterior and red from left to right. ''Right'': Medial representation of the 11 tracts: the cc (red), the CSTs (blue), the IFOs (yellow), the ILFs (green) the SLFs (purple) and UNCs (light blue).
 
!!Registration
This is the point, when we are going to use DTI-TK to register all the images to the common template. For this, you can refer to this [[Documentation.Registration|tutorial]], which walks you through different types of registration: rigid, affine and nonlinear. It also gives you three distinct commands depending on if you want to build a template while registering your data. In our case, we already have the template, so this is not necessary. The main commands to consider are
# rigid registration: dti\_rigid\_sn template.nii.gz subjs.txt EDS (subjs.txt is a text file containing the names of all the image files). Output: subject\_name.aff, a text file that contains an initialization matrix for the next step.
#affine registration: dti_affine_sn template.nii.gz subjs.txt EDS (keep the same subjs.txt). Output: subject_name_aff.nii.gz
# nonlinear registration: dti\_diffeomorphic_sn template.nii.gz subjs_aff.txt mask.nii.gz 6 0.002 (this time subjs_aff.txt consists of a list of all the affinely registered subjects'names). Output: subject_name_aff_diffeo.nii.gz

# dti_rigid_reg template.nii.gz tensor.nii.gz EDS 4 4 4 0.01
# dti_affine_reg template.nii.gz tensor.nii.gz EDS 4 4 4 0.01
# dti_diffeomorphic_reg template.nii.gz tensor_aff.nii.gz mask.nii.gz 1 6 0.002
 
and do so for each subject separately. After going through this section, all your subjects will be in a  ommon space. Here again, there's no harm in checking the registered images to make sure everything is ok!
 
!!Computing the Mean of the Fractional Anistropy or the Apparent diffusion Coefficient

At line 22, the path tsa_model_path needs to be edited and must point to the location of the medial representations.
That being done, create a .txt file containing the names of the registered subjects' files, such as dti_volume_list.txt
One this is done, just run the command: tsa_sampling dti_volume_list.txt.
Note: If you do not work with a cluster or your system does not allow to `qsub' jobs, then line 25 of tsa_sampling.sh shall be changed as well.
Replace: `qsub -b y -cwd -o \$\{medial\_pref\}.log -e \$\{medial\_pref\}.err /home/huiz/tools/dtitk/bin/medialTensorField \$\{medial\} \$\{dti\_volume\} maxFA \$\{medial\_pref\}.maxFA.vtk'

with
 

!!Statistical Analysis  
Non-parametric permutation-based suprathreshold statistical analysis. 
For this, use the script called gender_tsa_stats (which needs to be modified depending on the threshold you want to test and your degree of freedom, which depends on the number of  subjects), together with surfcluster.maxFA.config.txt that gathers all the information on the subjects.
At the bottom of the file surfcluster.maxFA.config.txt, the 0 and the 1 correspond to the patients and healthy subjects (or the opposite, it doesn't matter). 
Given the list in subjs_diffeo.txt, you can determined where the 0 and the 1 have to go in surfcluster.maxFA.config.txt.

Here is the abstract from the very good paper published by Zhang et al. in the Proceedings of the 4'^th^' International Conference on Biomedical Image registration in 2010 (see the paper [Publications.Publications|here]])
 
''Voxel-based analysis, either whole-brain or tract-specific, is a widely used approach for localizing white matter (WM) differences across populations using diffusion tensor imaging (DTI). A prerequisite to this approach is to spatially normalize all the subjects to a common template. The accuracy of spatial normalization can be improved by using a population-specific template that is, morphologically, most similar to the subjects in the population of interest. Here, we report the development of a population specific DTI template for the elderly using the publicly available IXI brain database [...]. The present template captures the average shape and diffusion properties of a population and contains segmentations of major WM fasciculi parcellated via fiber tractography. Furthermore, the segmentations are modeled using surface-based representation to support the tract-specific analysis recently proposed by Yushkevich et al. The template can be used to examine WM changes in neurodegenerative diseases and conditions.''
 
This abstract explains the core of the method and reflects upon another paper, by Yushkevich et al. which was published two years earlier \cite{yushkevich2008}. In this paper, the author built a children template.  From the two papers, you can easily guess the pipeline of the TSA technique. Here it is in a simplified version:
# Obtain an existing DTI template (they are not freely available for download yet, but will be soon. In the meantime, please send an email to cbrun@picsl.upenn.edu). Each template consists of a diffusion tensor atlas as well as medial representations of 11 tracts: the corpus callosum (CC), the corticospinal tracts (CST), inferior fronto-occipital tracts (IFO), inferior longitudinal tracts (ILF), superior longitudinal tracts (SLF), and uncinates (UNC)
Two templates have been built:
*an adult template, built with 78 subjects, 40 males/38 females (mean age: 39.5 pm 12), not yet published
* an aging template, generated from 51 subjects, 21 males/30 females (mean age: 70.1 pm 4), see (Zhang et al., 2010, [[Publications.Publications|here]]
 

# Register all the population diffusion tensor images to this template (rigid, affine and nonlinear registration)
# Sample the Fractional Anisotropy (or Apparent Diffusion Coefficient values) over all the registered subjects, compute the mean of this value at each point and project this value on the medial representation
# Perform a permutation-based non-parametric suprathreshold cluster analysis, to determine the WM clusters that are significantly different between your two populations.
 
 
!!Prior requirement
!!!Checking your DTIs 
Before any statistical analysis, or registration, you must have built your Diffusion Tensor images. For them to work seamlessly with DTI-TK, there are a few things you might want to check. 
# How were my images reconstructed? \url{http://groups.google.com/group/dtitk/web/data-format}. Here are a few tools that exist amongst others, which allow you to build your DTI from your DWI
# Then you want to make sure that the diffusivity units of the DT images are compatible with DTI-TK. Check this page to make sure (\url{http://groups.google.com/group/dtitk/web/important-information regarding-physical-unit-of-diffusivity})
# Last but not least, it is always a good idea (well, it's in fact highly recommended!) to visualize your images to make sure there is no outlier, which could critically influence your final analysis in a negative way. You will find a few very useful commands in this toolkit that will enable you to check these images \url{http://groups.google.com/group/dtitk/web/visualization-of-dti-volumes}. 
 
!!Obtaining the atlas
Here is an example of atlas you should have obtained.
%center%Attach:atlas_example_1.png
\caption{Left: Principal diffusion direction map of the DTI template. Blue represents the fibers that are going from inferior to superior regions, green posterior anterior and red from left to right. Right: Medial representation of the 11 tracts: the cc (red), the CSTs (blue), the IFOs (yellow), the ILFs (green) the SLFs (purple) and UNCs (light blue).}
  \end{center}
\end{figure}
 
\section{Registration}
This is the point, when we are going to use DTI-TK to register all the images to the common template. For this, you can refer to the tutorial \url{http://groups.google.com/group/dtitk/web/Registration%20of%20DTI%20Volumes}. This tutorial walks you through different types of registration: rigid, affine and nonlinear. It also gives you three distinct commands depending on if you want to build a template while registering your data. In our case, we already have the template, so this is not necessary. The main commands to consider are

\item rigid registration: dti\_rigid\_sn template.nii.gz subjs.txt EDS (subjs.txt is a text file containing the names of all the image files). Output: subject\_name.aff, a text file that contains an initialization matrix for the next step.
\item affine registration: dti\_affine\_sn template.nii.gz subjs.txt EDS (keep the same subjs.txt). Output: subject\_name\_aff.nii.gz
\item nonlinear registration: dti\_diffeomorphic\_sn template.nii.gz subjs\_aff.txt mask.nii.gz 6 0.002 (this time subjs\_aff.txt consists of a list of all the affinely registered subjects'names). Output: subject\_name\_aff\_diffeo.nii.gz

If you don't use cluster computing, then replace the above commands with