MNI tensor template

[Joe]

Does anyone know where to get an MNI space tensor template for use
with the dti_diffeomorphic_reg program? I've looked at the JHU
templates and they seem to provide individual images of an MNI FA map,
B0, and DWI, but not a tensor template required by the dti-tk
routine.

thanks
Joe

[hui Zhang]

Hi Joe,

The actual tensor image can be downloaded from UCLA LONI. You should
be able to find the exact link by google.

I do want to note that it does not work well with DTI-TK without a lot
of careful preproessing. What kind of studies are you interested in
doing? I may be able to suggest you better alternatives?

Gary

[KH]

Hi Gray

I also want to ask the question about LONI ICBM DTI-81 template.
After reading the "Use ICBM-DTI-81 Template for Registration & Spatial
Normalization " section, I think LONI DTI template isn't very suitable
to use in your tensor normalization pipeline.Because LONI DTI template
were constructed by averaging of all DWIs and used this average DWI
volume to perform further registration procedure without using any
tensor information. I followed the full procedure that you described
in "Use ICBM-DTI-81 Template for Registration & Spatial Normalization
" section and used "ellipsoidImageColOriAniScl" function to check the
orientation of each ellipsoid, i found some strange orientations in
corpus callosum and some fiber bundles.

Do you have another alternative suggestion to transform population
based 4space to MNI based template space ? I think these kind of
transformation is important , because some voxel based analysis
studies were required in reporting results in MNI space.

Best

KH

On 3月9日, 下午5時41分, hui Zhang <garyhuizh...@gmail.com> wrote:
> Hi Joe,
>
> The actual tensor image can be downloaded from UCLA LONI. You should
> be able to find the exact link by google.
>
> I do want to note that it does not work well with DTI-TK without a lot
> of careful preproessing. What kind of studies are you interested in
> doing?  I may be able to suggest you better alternatives?
>
> Gary
>

[Hui Zhang]

Hi  KH,

My apology for the late reply.

I also want to ask the question about LONI ICBM DTI-81 template.
After reading the "Use ICBM-DTI-81 Template for Registration & Spatial
Normalization " section, I think LONI DTI template isn't very suitable
to use in your tensor normalization pipeline.Because LONI DTI template
were constructed by averaging of all DWIs and used this average DWI
volume to perform further registration procedure without using any
tensor information. I followed the full procedure that you described
in "Use ICBM-DTI-81 Template for Registration & Spatial Normalization
" section and used "ellipsoidImageColOriAniScl" function to check the
orientation of each ellipsoid, i found some strange orientations in
corpus callosum and some fiber bundles.

The page that you are referring to has been in a "permanent" incomplete state.  In part, it is because the template is not particularly suited for DTI-TK for precisely the reasons you mentioned.  It will take some fiddling to get the brain extracted properly.  It's a shame that the license agreement for using the template prohibits me distributing any derivative content.

However, I'd be interested in seeing some screenshots of the orientation problem that you observed.  The ellipsoid tool has a built-in keystroke ('p') to dump a screenshot.

 

Do you have another alternative suggestion to transform population
based 4space to MNI based template space ? I think these kind of
transformation is important ,  because  some voxel based analysis
studies were required in reporting results in MNI space.

You can use the ICBM template successfully for mapping a local template to the standard space with affine only registration as described here

http://groups.google.com/group/dtitk/web/Image%20Mapping%20to%20Standard%20Space

Just skip the nonlinear portion.  The ICBM template, like other MNI template, is produced with only affine alignment anyway.  The very coarse anatomical detail in the template can only lead to spurious correspondences that are not anatomically meaningful.

Hope this helps,

Gary

[Gustavo Sudre]

Gary,

I'd like to use MNI coordinates to create ROIs for a tract-specific analysis, instead of manually drawing them. The link described in your last post looks relevant, but it's broken. Could you please post again how to go from a local template to a template in MNI space?

Thanks,

Gustavo

[Hui Zhang]

Hi Gustavo,

You are right.  That link was phased out by Google.  Here is the corresponding link in the current site:

http://dti-tk.sourceforge.net/pmwiki/pmwiki.php?n=Documentation.StdSpace

Gary

--
 
---
You received this message because you are subscribed to the Google Groups "DTI-TK" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dtitk+un...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[Gustavo Sudre]

Gary,

Thanks for the link. Now, I'm having difficulties converting the ROIs, defined in MNI space, to the dimensions used in DTI-TK. For example, say I draw MNI-guided ROIs using IIT3mean_t1.nii.gz or even a Tailarached version of IIT3mean_t1_256.nii.gz. The dimension of the final result don't match 256x256x256 expected by the other files in the template (e.g. IIT3mean_tensor_256_reorient.nii.gz). 

I tried resampling them with TVresample, but I get:

TVResample -in IIT3mean_t1_256_tlrc.nii.gz -size 256 256 256

interpolation option is LEI

Reading anat_tlrc.nii.gz ... Warning: Volume uses unexpected NIfTI intent code

Done in 0.079768s

Segmentation fault: 11

Do you know what's going on? Or, even better, do you have different guidelines for using MNI-defined ROIs when creating ROIs for the tract-speficifc analysis? The background reason here is to use MNI coordinates to better inform the drawing of ROIs, such as the example in http://dti-tk.sourceforge.net/pmwiki/pmwiki.php?n=Documentation.TsaAdvanced

Thanks!

Gustavo 

You received this message because you are subscribed to a topic in the Google Groups "DTI-TK" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/dtitk/LUYVjSAkTBw/unsubscribe.
To unsubscribe from this group and all its topics, send an email to dtitk+un...@googlegroups.com.

[Hui Zhang]

Hi Gustavo,

The problem here is quite simple actually.  TVResample is for resampling tensor volumes, hence the prefix "TV".  It only works with tensor volumes that have six components only.

To define ROIs, you can propagate them from a MNI space to your local template through registration as described here:

http://dti-tk.sourceforge.net/pmwiki/pmwiki.php?n=Documentation.OptionspostReg

Gary

[Gustavo Sudre]

Gary,

Thanks for getting back to me. It turns out hat I was able to propagate my ROI, but now I'm getting stuck in the last part of the tutorial to create our own ROI, the one that requires cmrep_fit. I saw in the other thread that you have already contacted Paul about it, so I'll wait to see if he can figure out what the problem is.

Thanks,

Gustavo