---

From: "lzollei@nmr.mgh.harvard.edu" <lzollei@nmr.mgh.harvard.edu>
To: Antonella Kis <atorok9@yahoo.com>
Cc: Lilla Zollei <lzollei@nmr.mgh.harvard.edu>
Sent: Friday, November 18, 2011 2:11 PM
Subject: Re: cvs_avg35

Hi Antonella,

> I know that mri_vol2vol just applies the transform. --no-save-reg just
> tells the
> command not to save a copy of the inverse transform that gets applied. My
> question is do I need to run the mri_vol2vol with the fla --no-save-reg 
> or with the flag --save-reg if I want to resample the diffusion data of
> one subject  in the structural space and same later in the CVS space? Do I
> need to save the invers transform or not?

That flag is not important at all for your purposes. We suggest using
--no-save-reg in order to avoid having multiple copies of files around.
There is no --save-reg, by the way, as that is the default behavior.

--Lilla

> ________________________________
>  From: Lilla Zollei <lzollei@nmr.mgh.harvard.edu>
> To: Antonella Kis <atorok9@yahoo.com>
> Cc: "freesurfer@nmr.mgh.harvard.edu" <freesurfer@nmr.mgh.harvard.edu>
> Sent: Thursday, November 17, 2011 4:40 PM
> Subject: RE: cvs_avg35
>
>
> Hi Antonella,
>
>> 1) Can you please help me find or get the the cvs_avg35/mri/norm.mgz? I
>> saw online the same question was posted on June 2011 but there is no
>> answer for the question online:
>>
>> http://www.opensubscriber.com/message/freesurfer@nmr.mgh.harvard.edu/15506607.html
>
> Have you run the cvs_registration code? If you have, then you have to have
> the template. If you have not, I can make the template subject available
> for you for download.
>
>> 2) Since for my TBSS TBSS study was used by the TBSS the  nonlinear
>> registration  FNIRT (is about 15 times faster than IRTK on FA data and
>> possibly a little more 'accurate') I would like to use for a Tal space a
>> non-linear talairach registration. Can you please give
>> me some directions on  how I can do this? What is the default Tal
>> registration (linear or non-linear) for dt_recon?
>
> dt_recon produces a linear registration to the Tal space. If you have ran
> recon-all on your structural data you should also have access to
> talairach.m3z that aligns your subject in the fsaverage space.
>
>> 3) When dt_recon is  aligning  the diffusion data in the Tal space is in
>> fact aligning the register.dat data (difussion data already aligned to
>> the structural data) and the output is the fa-tal.nii.gz? If not why I
>> need the register.dat file?
>
> I don't understand your question, but
> * register.dat is the registrationfile that encodes the transform going
> from the diffusion to the structural space. If you want to resample any of
> your diffusion related files in the structural space you will need to use
> mri_vol2vol and that file to do so.
> * fa-tal.nii.gz is a completely different file. It is the fa volume that
> got resampled in the Tal space using a linear transformation
> (talairach.xfm via the structural data).
>
>> 4) So by my understanding each subject will have a file called
>> fa-tal.nii which in fact represents the subject's registration on the
>> Tal-space? How about the mean diffusivity? I don't have only one file
>> for this adc.nii. What if I want to do a non_FA values study on
>> the tal space?
>
> You need to run the follwoing command, replacing fa with the scalar volume
> of your interest.
>
> set reg = register.dat
> set fa = fa.nii
> set fatal = fa-tal.nii.gz
> set cmd = (mri_vol2vol --reg $reg --tal --mov $fa --o $fatal)
>
>> 5) I understood that the 'registration.dat' file describes a rigid-body
>> transformation from the diffusion to the structural space and in order 
>> to resample the structural volume in the diffusion space, I will need to
>> apply the inverse of this transformation to my
>> input volumes and to use for this the 'mri_vol2vol':
>>
>> mri_vol2vol --mov $TUTORIAL_DIR/$subj/dtrecon/lowb.nii \
>>             --targ $SUBJECTS_DIR/$subj/mri/wmparc.mgz \
>>             --inv --interp nearest --o
>> $SUBJECTS_DIR/$subj/mri/wmparc2diff.mgz \
>>             --reg $TUTORIAL_DIR/$subj/dtrecon/register.dat --no-save-reg
>> Is the register.dat file created by this function mri_vol2vol?  I
>> believe the registration between the diffusion and structural space (so
>> the register.dat file) has already
>> been computed as part of the previous pre-processing step so while using
>> the mri_vol2vol as above do I need to use the '--no-save-reg' flag as in
>> the above example or the '--save-reg' flag?
>
> register.dat gets created using dt_recon and mri_vol2vol just applies that
> transform. --no-save-reg just tells the command not to save a copy of the
> inverse transform that gets applied.
>
>> 6) To resample the diffusion data of one subject  in the CVS space do I
>> need to run after dt_recon the mri_vol2vol as below?
>> mri_vol2vol --targ
>>  $FREESURFER_HOME/subjects/cvs_avg35/mri/norm.mgz \
>>             --m3z
>> $SUBJECTS_DIR/$subj/cvs/combined_tocvs_avg35_elreg_afteraseg-norm.m3z \
>>             --noDefM3zPath --reg
>> $TUTORIAL_DIR/$subj/dtrecon/register.dat \
>>             --mov $TUTORIAL_DIR/$subj/dtrecon/fa-masked.mgz \
>>             --o
>> $TUTORIAL_DIR/$subj/dtrecon/fa-masked.ANAT+CVS-to-avg35.mgz \
>>             --interp trilin --no-save-reg
>> Do I need to run the mri_cvs_registration on the structural data in
>> order to get the above .m3z file
>> 'combined_tocvs_avg35_elreg_afteraseg-norm.m3z' before I run the
>> mri_vol2vol?
>> If I want to have my second registration step done in the Tal space not
>> CVS space since after running dt_recon I am getting the fa-tal.nii.gz
>> and the fa-masked.mgz
>> is this the equivalent file  to  fa-masked.ANAT+CVS-to-avg35.mgz (if I
>> use the CVS space)? In this case I don't need to run the last
>> mri_vol2vol command?
>> Sorry for asking so many questions but it is very confusing the online
>> material.
>>
> Yes, you will run mri_cvs_register on the structural data and that will
> compute your m3z file.
>
> fa-masked is in the native diffusion space and fa-tal is in the Talairach
> space so they are not equivalent with each other, they don't live in the
> same space.
>
> --Lilla
>
>> From: Lilla Zollei <lzollei@nmr.mgh.harvard.edu>
>> To: Antonella Kis <atorok9@yahoo.com>
>> Cc: "freesurfer@nmr.mgh.harvard.edu" <freesurfer@nmr.mgh.harvard.edu>
>> Sent: Tuesday, November 15, 2011 5:11 PM
>> Subject: RE:[Freesurfer] dt_recon registration
>>
>>
>> Hi Antonella,
>>
>> > 1) Why I don't have in my $FREESURFER_HOME/subjects the
>> cvs_avg35/mri/norm.mgz? In my subjects directory I only have my
>> subjects and nothing else. How can I get the  cvs_avg35/mri/norm.mgz?
>>
>> $FREESURFER_HOME is different from your $SUBJECTS_DIR. Could you double
>> check?
>>
>> > 2) Do you believe that for my subjects-all children ages 6-13 the
>> results are gone be OK if I use the talairach registration and
>> re-sampling? As I mentioned in my previous email I would like to do
>> > a two-group study similar with TBSS so to do a statistical analysis on
>> the FA and non_FA data such as mean diffusivity, to see if there are
>> any clusters that may survive by multiple comparisons without any
>> covariant.
>>
>> Are you referring to a linear or a non-linear talairach registration? It
>> will be you who needs to decide which space to use for analysis. You
>> would want all your data to be well aligned in that space in order to
>> achieve a meaningful group comparison. The decision
>> might also depend on in which space you want to report your final
>> results in.
>>
>> > 3) Do you have any suggestion regarding which registration is better
>> to be used for this study?
>>
>> I am not familiar with your study, so it would be hard to say.
>>
>> > 4) How I can check for each individual subject  if the registration
>> worked out well if for example  I choose the talairach space?
>> > Do I run:
>> >
>> > tkregister2 --mov lowb.nii --reg register.dat --surf
>> > or which file is in fact the registered data file? Is the
>> fa-tal.nii.gz or the lowb.nii.gz? You mentioned that the default
>> registration is on  talairch space and the I see on dt_recon tutorial
>> >  that is also doing a registration of lowb to same-subject anatomical
>> so I have two registration done when I run the dt_recon?
>>
>> dt_recon has two registration steps. One aligns the diffusion data
>> (lowb) to the structural (registration.dat is the default name of the
>> registration file) and the other aligns the diffusion data in the Tal
>> space (the resampled fa file in that space is fa-tal). You
>> can use both freeview and thregister to check the alignment of your
>> volumes.
>>
>> Lilla
>>
>>
>>
>>
>> >__________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________
>> _
>> > From: Lilla Zollei <lzollei@nmr.mgh.harvard.edu>
>> > To: Antonella Kis <atorok9@yahoo.com>
>> > Cc: "freesurfer@nmr.mgh.harvard.edu" <freesurfer@nmr.mgh.harvard.edu>
>> > Sent: Monday, November 14, 2011 4:13 PM
>> > Subject: Re: [Freesurfer] dt_recon registration
>> >
>> >
>> > Hi Antonella,
>> >
>> > > 1) I will be very grateful if you can confirm if my coding is write.
>> First can you please let me know if by default dt_recon is re-sample
>> FA to talairch space? In this case because I want to register   on
>> CVS should I add the  --no-tal flag? I am asking this
>> > question
>> > > because after I ran the following command I see in my  output
>> directory called P07_dti a file called fa_tal.nii as well as
>> fa_tal.nii.reg.
>> >
>> > The talairach registration and re-sampling does take place by default,
>> but that should not affect your further processing as the fa.nii file
>> (in the native space) is not deleted from the directory. So you can
>> either leave the default settings or you can also use
>> > --no-tal.
>> >
>> > > 2) Do I need this files if I want to do a group study for my FA and
>> mean diffusivity, similar to TBSS in FSL?
>> >
>> > Are you referring to the tal registration files? If you want to do the
>> group study in the CVS spacem then you will not need the tal
>> registration files.
>> >
>> > > 3) Can I use the talairch registration  or should I use the CVS?
>> >
>> > Depends on where you want to do your analysis.
>> >
>> > > 4) Should I add the flag  --no-tal?
>> >
>> > You can. If you don't you just end up with two additional files.
>> >
>> > > This is what I was running:
>> > >
>> > > setenv FREESURFER_HOME /usr/local/freesurfer
>> > > setenv SUBJECTS_DIR $FREESURFER_HOME/subjects/DTI_FS/dtrecon
>> > >
>> > > dt_recon --i P07_dti.nii.gz  -- b P07.bvals P07.bvects --s P07 --o
>> P07_dti  where
>> > >
>> > > P07_dti.nii.gz is my DTI volume
>> > > P07.bvals P07.bvects are the bvals and bvects for my subject
>> > > P07 is my structural recons data for my P07 subject
>> > >
>> > >
>> > >
>> > > 5) From the dt_recon --help I understand that dt_recon is doing the
>> registration of lowb to same-subject anatomical using
>> > >    FSLs flirt (creates mask.nii and register.dat). 
>> > >
>> > > mri_vol2vol --mov
>> /usr/local/freesurfer/subjects/DTI_FS/dtrecon/P07_dti/lowb.nii \
>> > >             --targ $SUBJECTS_DIR/P07/mri/wmparc.mgz \
>> > >             --inv --interp nearest --o
>> $SUBJECTS_DIR/P07/mri/wmparc2diff.mgz \
>> > >             --reg
>> /usr/local/freesurfer/subjects/DTI_FS/dtrecon/P07_dti/register.dat
>> --no-save-reg
>> >
>> > Actually dt-recon now uses bbregister, but the output of this step is
>> still register.dat. With the above you move the wmparc into the
>> diffusion space.
>> >
>> > > 6) Should I run mri_vol2vol  to re-sample the structural volume in
>> the diffusion space and register in the CVS space:
>> > >
>> > >
>> > > mri_vol2vol --targ $FREESURFER_HOME/subjects/cvs_avg35/mri/norm.mgz
>> \
>> > >             --m3z
>> $SUBJECTS_DIR/$subj/cvs/combined_tocvs_avg35_elreg_afteraseg-norm.m3z
>> \
>> > >             --noDefM3zPath --reg
>> $TUTORIAL_DIR/$subj/dtrecon/register.dat \
>> > >             --mov
>> /usr/local/freesurfer/subjects/DTI_FS/dtrecon/dti_15_p7/fa-masked.mgz
>> \
>> > >             --o
>> /usr/local/freesurfer/subjects/DTI_FS/dtrecon/dti_15_p7/fa-masked.ANAT+CVS-to-avg35.mgz
>> \
>> > >             --interp trilin --no-save-reg
>> >
>> > With the above you are going to move your diffusion file (fa) into the
>> CVS space (by implicitely combining a moprh moving the diffusion file
>> to the structural space and then to the CVS space) and not resampling
>> the structural volume in the diffusion space.
>> >
>> > > 7) I found online that the default CVS template is cvs_avg35 which
>> is supplied with my FreeSurfer distribution. Can you please tell me
>> in which directory is located or how I can obtain
>> > > the cvs_avg35?
>> >
>> > It should be in $FREESURFER_HOME/subjects (as indicated above in your
>> command --targ).
>> >
>> > --Lilla
>> > _______________________________________________
>> > Freesurfer mailing list
>> > Freesurfer@nmr.mgh.harvard.edu
>> > https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
>> >
>> >
>> > The information in this e-mail is intended only for the person to whom
>> it is
>> > addressed. If you believe this e-mail was sent to you in error and the
>> e-mail
>> > contains patient information, please contact the Partners Compliance
>> HelpLine at
>> > http://www.partners.org/complianceline . If the e-mail was sent to you
>> in error
>> > but does not contain patient information, please contact the sender
>> and properly
>> > dispose of the e-mail.
>> >
>> >
>> >
>> >
>> >
>> >
>>
>>
>>
>>
>>