[caret-users] Mapping BrainVoyager statistical volumes to PALS atlas

[Manny Vindiola]

Donna,

Thanks for your reply. I to am hoping that I can massage the BV files
so that I can use the AFNI surface as a proxy. Here is some
information regarding the functional and statistical formats BV uses
for a talairached dataset.

The data sits in a bounding box around the anatomical with dimensions:
139 x 175 x 121

According to one of the mapper help files the AFNI talairach grid is:
161x191x151

I don't know if there is a straight forward way to adjust the BV files
to enlarge the bounding box. Obviously I can zeropad the file but I am
not sure that will have the appropriate effects. Would it help if I
uploaded an anatomical and functional image for you somewhere?

thanks for your help,
Manny

> Hi Manny,
>
> I have no Brainvoyager experience, but I map NIfTI to PALS often.  Which
> version of Caret are you running?  John Harwell fixed some NIfTI issues
> a while back, but I don't recall exactly when.  If you are using the
> most recent version (Version 5.61 - 12 November 2008), then those
> changes should be incorporated already.
>
> The volume issue must be resolved before the next issue can be
> pinpointed:  Why isn't the AFNI surface a good proxy for Brainvoyager,
> since both packages use the Talairach & Tournoux 1988 (T&T88) recipe for
> spatial normalization?  Or at least they used to; I'm guessing most AFNI
> users now use @auto_tlrc to spatially normalize their subjects.  While
> it still uses the T&88 bounding box / scale, it uses an affine transform
> to get there, rather than the piecewise linear process defined in
> T&T88.  I do not know whether Brainvoyager now offers an affine option;
> however, the PALS AFNI surfaces were normalized using the old, manual
> method, because the affine option did not exist when these surfaces were
> created.
>
> If it is really true that the PALS AFNI surfaces are poor proxies for
> Brainvoyager, then we will need to consider developing Brainvoyager
> surfaces.  But I would like to see some evidence before adding that task
> to our queue.
>
> Donna