Donna, the link appears to be password protected, thank you Traci 

http://brainmap.wustl.edu/pub/donna/FREESURFER/SCRIPTS/2009_09/

-----Original Message-----
From: [email protected] 
[mailto:[email protected]] On Behalf Of Donna Dierker
Sent: Thursday, October 08, 2009 1:42 PM
To: Caret, SureFit, and SuMS software users
Subject: Re: [caret-users] do spherical registration using freesurfer generated 
surfaces

Hi Traci,

No, I don't. *Someone* -- YOU KNOW WHO YOU ARE -- has me working on
similar scripts, but different enough to be both a fair amount of work
and not as applicable to what I had in mind. I hope to wrap that up this
week (but we'll see).

Meanwhile, I posted interim versions of what I do have:

http://brainmap.wustl.edu/pub/donna/FREESURFER/SCRIPTS/2009_09/

freesurfer2caret.Ver2a.preborder.mni2wustl.sh - imports Freesurfer to
Caret and creates inputs for next script
draw_core6_landmarks.tim.sh - draws borders
border_variability.sh - generates quantitative measures of landmark quality
generate_scene.check.sh - generates captures, scenes of borders
[manual step: Layers: Borders: Border Update]
postborder.sh - smooths medial wall, morphs and aligns sphere, registers

If you have more than a handful of subjects, it will probably be worth
getting these to work.

Realistically, I hope to have something applicable to more typical
Freesurfer input by 10/31.

The script above had more steps like this than the typical stream would
have, due to strangeness of the input data and inadequate talairach.xfm
(not imgreg's fault):

caret_command -surface-apply-transformation-matrix
$CASE.$HEM_FLAG.MIDTHICKNESS.coord $CASE.$HEM_FLAG.topo
$CASE.$HEM_FLAG.MIDTHICKNESS_orig.coord -matrix $MATRIX

Donna


On 10/08/2009 01:15 PM, Sandoval, Traci I wrote:
> Hi Donna, do you have these cool caret scripts ready ie...
>
> Next week, I'll adapt these import scripts to a dataset more like what yours 
> is likely to be. When I am done, I can post the resulting scripts.
>   
>> Can you ping me in two weeks?
>>
>> Donna
>>     
>
> Thanks for all your work and help. Traci
>
> Traci Sandoval
> Research Assistant
> NeuroPsychometric Research Lab
> School of Behavioral and Brain Sciences - University of Texas at Dallas 
> Center for BrainHealth - University of Texas at Dallas
> Department of Psychiatry - University of Texas Southwestern Medical
> www.utdallas.edu/research/nprlab
> 972-883-3270
>
>
>
> -----Original Message-----
> From: [email protected] 
> [mailto:[email protected]] On Behalf Of Donna Dierker
> Sent: Friday, September 18, 2009 11:20 AM
> To: Caret, SureFit, and SuMS software users
> Subject: Re: [caret-users] do spherical registration using freesurfer 
> generated surfaces
>
> Traci,
>
> Yes, one lab here at Wash U is doing that. A few comments:
>
> * Since you don't have a flat map, the reference distance for the
> posterior sylvian starting point changes from 9mm posterior to the
> SF/SF2 intersection (relative to the flat map) to 12mm posterior to the
> SF/SF2 intersection on the spherical map.
>
> * It's tougher to discern the full sylvian trajectory on the spherical
> map, as well as center the sylvian on the main window, such that you can
> see the whole length. Fortunately, the absolutely fun-to-use Layers:
> Borders: Border Update makes this easy to fix, if you can't draw the
> whole length in one go. (If your older Caret version doesn't have this
> feature, you really MUST upgrade.)
>
> * You have to draw the medial wall dorsal and ventral in separate
> segments, and you have to pay attention to their orientation (dorsal
> starts anteriorly, ventral posteriorly). Calcarine starts posterior.
> Flattening took care of this for you.
>
> * The lack of a compressed medial wall view makes it harder to draw the
> medial wall ventral.
>
> Donna
>
> On 09/18/2009 10:19 AM, Sandoval, Traci I wrote:
>   
>> Donna, That’s sounds great, I have scripts that take my freesurfer brains 
>> all the way to flattening in caret, but I flatten and draw borders manually. 
>> I noticed you mentioned you guys are doing the auto-landmarks on the 
>> spherical surface, and not flattening at all. Can I do that manually as 
>> well? And skip flattening? Thanks Traci
>>
>> -----Original Message-----
>> From: [email protected] 
>> [mailto:[email protected]] On Behalf Of Donna Dierker
>> Sent: Friday, September 18, 2009 9:26 AM
>> To: Caret, SureFit, and SuMS software users
>> Subject: Re: [caret-users] do spherical registration using freesurfer 
>> generated surfaces
>>
>> Traci,
>>
>> I do have some scripts I could make available now, but in two-four
>> weeks, I'll have more general purpose ones. The most complete set were
>> actually developed for a data set segmented in Caret. But they include
>> the auto-landmarks, border variability (QA), generating scenes/captures
>> showing the borders, and running registration via command line. There
>> also are several versions of a "preborder" script that covers importing
>> from Freesurfer, but most of these used anatomical data that was in
>> 711-2* space before feeding to Freesurfer. Your data is unlikely to fit
>> that category.
>>
>> Next week, I'll adapt these import scripts to a dataset more like what
>> yours is likely to be. When I am done, I can post the resulting scripts.
>>
>> Can you ping me in two weeks?
>>
>> Donna
>>
>> On 09/17/2009 04:07 PM, Sandoval, Traci I wrote:
>>   
>>     
>>> Donna, I would love to use these scripts that aren't publicized yet. How 
>>> can I attain said scripts?!! Thank you for all your work and questions 
>>> answered. Traci
>>>
>>>
>>> Traci Sandoval
>>> Research Assistant
>>> NeuroPsychometric Research Lab
>>> School of Behavioral and Brain Sciences - University of Texas at Dallas 
>>> Center for BrainHealth - University of Texas at Dallas
>>> Department of Psychiatry - University of Texas Southwestern Medical
>>> www.utdallas.edu/research/nprlab
>>> 972-883-3270
>>>
>>>
>>>
>>> -----Original Message-----
>>> From: [email protected] 
>>> [mailto:[email protected]] On Behalf Of Donna Dierker
>>> Sent: Thursday, September 17, 2009 8:44 AM
>>> To: Caret, SureFit, and SuMS software users
>>> Subject: Re: [caret-users] do spherical registration using freesurfer 
>>> generated surfaces
>>>
>>> Hi Jidan,
>>>
>>> See inline replies below.
>>>
>>> Donna
>>>
>>> On 09/17/2009 02:06 AM, z丹丹 wrote:
>>>   
>>>     
>>>       
>>>> Hi,
>>>>
>>>> I already have fiducial surfaces generated by Freesurfer, which are
>>>> all in the MNI space. I want to do spherical registration to an atlas
>>>> spherical surface. The steps i need to do are:
>>>>
>>>> 1, generate the ellipsoid, spherical surfaces, and flat maps.
>>>> 2, create the 6 landmarks on both individual and atlas surface.
>>>> 3, do spherical mapping.
>>>>
>>>> Are the procedure right? Thanks! 
>>>>     
>>>>       
>>>>         
>>> Funny you should ask. We (the Van Essen Lab) are using Freesurfer for no
>>> less than five projects, and we have shell scripts that streamline much
>>> of the import, border drawing, QA, and registration. There is a feature
>>> we call "auto-landmarks" that draws the borders reasonably well. Many
>>> borders need touch-ups, but a single rater fixed 72 hemispheres' borders
>>> in 1.5 days. If you have only a handful of subjects, it's not clear
>>> whether it would be worth it for you to use our methods. But if you have
>>> dozens, then they would save you a lot of time.
>>>
>>> We haven't broadly publicized these features/scripts, because they're
>>> still in the final stages of development and testing. But they have been
>>> working well for us.
>>>
>>> Even if you don't use auto landmarks or these scripts, one question
>>> stands out in your steps above: Do you need a flat map? No, not really.
>>> It certainly is easier to draw landmarks on the flat map than the
>>> sphere; however, I'm not sure it's worth the hassle. Then again, if
>>> you're not using the scripts, it might be the shortest path to registration.
>>>   
>>>     
>>>       
>>>> In fact before I move to the 2nd step, I was stuck in the flat map
>>>> generation. I met two problems:
>>>> 1, As my surface is in MNI space, the surface orientation is not right
>>>> in the CARET software. I found that the anterior and posterior are
>>>> reversed with superior and inferior. Is there any method to roate it
>>>> into the right orientation in CARET?
>>>>     
>>>>       
>>>>         
>>> Yes, but this is not consistent with my experience of MNI space nor
>>> Freesurfer surface orientation. While MNI volumes are sometimes stored
>>> using left-handed coordinate systems (what some call radiological
>>> orientation), the coordinates are always reported with -x being left, +x
>>> being right; -y being posterior, +y being anterior; -z being inferior,
>>> +z being superior. I've never seen a Freesurfer surface with the Y and Z
>>> axes flipped -- not even ones written in MNI space.
>>>
>>> It's easy to flip a surface from LAS to LPI
>>> (http://brainmap.wustl.edu/OLD/SureFit/orient.html) like this:
>>>
>>> caret_command -surface-apply-transformation-matrix $COORD $TOPO
>>> $FLIPPED_COORD -matrix "1 0 0 0 0 -1 0 0 0 0 -1 0 0 0 0 1"
>>>
>>> But I really would not brush this off. I would get to the bottom of why
>>> the supposedly MNI surface is not in the expected orientation. The
>>> evidence suggests it is not in MNI space, as asserted.
>>>   
>>>     
>>>       
>>>> 2, The other problem is , I found that the origin of my surface is not
>>>> AC. So when I flatten the surface, it can't find the correct medial
>>>> wall to cut. Do you think there is a way to solve this too?
>>>>     
>>>>       
>>>>         
>>> This is more evidence that the surface is not really in MNI space. How
>>> did it get transformed to MNI?
>>>
>>> Note that the MNI origin is not exactly the AC; see
>>> http://imaging.mrc-cbu.cam.ac.uk/imaging/MniTalairach for details. But
>>> it is certainly close enough to the AC that the compressed medial wall
>>> view should be centered over the medial wall. The fact that it is not
>>> indicates a problem with your transformation to MNI space.
>>>   
>>>     
>>>       
>>>> Thanks a lot!
>>>>
>>>> Jidan
>>>>     
>>>>         

_______________________________________________
caret-users mailing list
[email protected]
http://brainvis.wustl.edu/mailman/listinfo/caret-users
_______________________________________________
caret-users mailing list
[email protected]
http://brainvis.wustl.edu/mailman/listinfo/caret-users

Reply via email to