[caret-users] Talairach group data to individual surface

[Donna Dierker]

Hi Sarah,

Problems like these are relatively common -- especially if the surface 
is old (i.e., generated by SureFit, which used to place the origin at 
the lower left corner of the cropped bounding box -- not the anterior 
commissure or atlas space origin, as Caret commonly does).

I would make a FAQ, but the problems tend to be so case-specific, that 
common threads are few. Uploading the dataset and letting me have a look 
is usually the shortest path to solution:

http://pulvinar.wustl.edu/cgi-bin/upload.cgi

I need the following files:

* fiducial surface coord
* topo file
* anatomy volume to which functional aligns
* functional volume (optional -- almost always, if I can get surf and 
anat aligned, the rest follows)

If your surface is in VTK form, that can substitute for coord/topo.

Donna

On 12/10/2007 02:04 PM, Sarah Hillenbrand wrote:
> Hi,
>
> Thanks for the advice Donna.  That is, in fact, what we are trying to do.
> However, I am realizing the source of my problems is at a much more basic
> level, so I have abandoned the talairached data and am working with
> individual data.
>
> Right now, I'm attempting to map some individual data onto the individual
> surface and am running into some problems.
>
> I wrote the functional data into the original anatomical's grid as per
> http://brainmap.wustl.edu/help/mapper.html#afniwriteanat.
>
> I then proceeded to mimic the directions at
> http://brainmap.wustl.edu/help/map_individual.html
> Using my functional data in anatomical grid, the original anatomical to
> which it was mapped, and a surface I generated for this same subject using 4
> anatomicals (including the aforementioned "original" one),
> registered/averaged/uniformized/resampled/converted to MINC in  AFNI.
>
> The functional data maps fine onto the original anatomical, but not onto the
> surface.  I think there may be some alignment issues between the original
> anatomical and the processed one used for the segmentation/surface.  Can you
> recommend something to fix this?
>
> Thanks so much,
> Sarah
>
>
>
>
>
> On 12/6/07 10:32 AM, "Donna Dierker" <donna at brainvis.wustl.edu> wrote:
>
>   
>> Hi Sarah,
>>
>> This is not something we do every day -- map group results to an
>> individual's surface.  Maybe some context might clarify.
>>
>> But let me illustrate an example where something like this might make
>> sense.  Suppose I have functional group differences, and I want to
>> investigate whether anatomical difference might underly the functional
>> differences.  I can imagine doing something like this:
>>
>> * Use some sort of volume-based registration (VBR) to warp the
>> individual's anatomical volume into the same stereotaxic space as my
>> functional group results.  Spend a fair amount of time fussing about
>> what VBR method to use (e.g., linear vs nonlinear).  It might be
>> reasonable to use the same VBR method used to generate the functional
>> group results, but give this step some thought before charging ahead.
>>
>> * Use caret_command or AFNI's Vecwarp to apply the resulting affine
>> transform, if a linear VBR method was used, to your individual's
>> surface, so the surface is now in atlas space.  (If a nonlinear method
>> is used, then you'll need to apply the deformation map using alternate
>> methods.  The SPM links here give you a feel for how that works:
>>
>>     http://brainvis.wustl.edu/help/pals_volume_normalization
>>
>> * Map the functional results onto the atlas-space surface.
>>
>> * Perhaps smooth/threshold the results, using Caret's Attributes:
>> Metric: Clustering and Smoothing and/or Surface: ROI features.
>>
>> * Perhaps even use Surface: ROI to threshold the map and convert it to
>> Paint, so that now I have a ROI on the individual's surface; then,
>> Attributes: Paint: convert paint column to paint volume to project the
>> ROI on the individual's surface to the atlas-warped individual's
>> anatomy.  I might use this as a starter segmentation, perhaps clean it
>> up, perhaps not, for the purpose of applying some sort of volumetric
>> morphometric method.
>>
>> This may be nothing like what you want to do, but it's the kind of
>> context under which it might make sense to map group fMRI onto
>> individual surfaces.
>>
>> Donna
>>
>> On 12/06/2007 09:11 AM, Sarah Hillenbrand wrote:
>>     
>>> I have functional group data in Talairach space that I need mapped to an
>>> individual surface.  I'm having trouble with this and am not sure at which
>>> step in the process I may have gone astray.
>>>
>>> What are the correct steps to follow here?
>>>
>>> Much obliged,
>>> Sarah
>>>
>>> _______________________________________________
>>> caret-users mailing list
>>> caret-users at brainvis.wustl.edu
>>> http://pulvinar.wustl.edu/mailman/listinfo/caret-users
>>>
>>>   
>>>       
>> _______________________________________________
>> caret-users mailing list
>> caret-users at brainvis.wustl.edu
>> http://pulvinar.wustl.edu/mailman/listinfo/caret-users
>>     
>
> _______________________________________________
> caret-users mailing list
> caret-users at brainvis.wustl.edu
> http://pulvinar.wustl.edu/mailman/listinfo/caret-users
>
>