I was just thinking FSL's command line utilities also could get the job done, but again -- don't think there's a Windows version. (Morgan would certainly know.;-) My guess is John will need to go the Cygwin route, unless he gets MacOSX or Linux access soon.

To make John feel a bit worse, note the "Unix" tag way at the bottom of the MRIcro page (http://www.sph.sc.edu/comd/rorden/mricro.html):

FormatConvert [Unix] Pittsburgh 1.0, Afni, GE (I*.nnn files, E*.MR files), BrainVoyager (VMR, VTC, STC, VMP)

Keep hounding Brain Innovation for real NIfTI support.

On 09/21/2006 04:03 AM, Morgan G. Hough wrote:

I don't know if someone has already suggested this but I have used
Format Convert from the Fiswidget tools to convert BrainVoyager data to
NIfTI but I think you will still need some form of UNIX environment (or
Cygwin) to get this working. Here's a link to check out.

Hope this helps.


On Thu, 2006-09-21 at 09:09 +0100, John Taylor wrote:
Donna - Thanks for the ideas. You confirmed exactly what I saw in terms of the memory stick up. The BVQX user guide does imply the existence of a plugin to provide a NIfTI export option but I have so far been unable to trace the actual plugin and as yet have had no reply from other BV users regarding this option. If there is anyone on this list who has sucessfully exported tal-space .vmr files into CARET as NIfTI, I'd be very interested to hear from you! I currently only have access to a Windows PC in the lab (although this is likely to change) which rules out AFNI options for now, so I may need to hold fire on this. I'll look at your other suggestion today.



Donna Dierker wrote:

Hi John,

On 09/19/2006 05:24 AM, [EMAIL PROTECTED] wrote:

Of course, I meant 'Radiological' rather than 'Neurological'! Apologies for any
... and previously wrote:

On 09/19/2006 05:19 AM, [EMAIL PROTECTED] wrote:

When opening new .hdr data files exported from Brain Voyager QX, the volumes require reorienting from neurological > LPI. I am using the 'Orientation' tab in the 'Volume attributes' editor to do this. However, although I am able to flip around the X and Y axes without problems, both the 'Rotate Clockwise' and 'Convert to LPI' commands appear to introduce an error such that the 'Resample' tab displays voxel dimensions of 0.000. If I then attempt to resample to 1 x 1 x 1, Caret crashes. Am I approaching the reorientation incorrectly or have
others experienced the same problem?
Any help appreciated

John Taylor
I was able to reproduce this behavior, so it seems like something John should look at. Even after changing the Resample x spacing from 0.0 to 1.0, I still get an OUT OF MEMORY stickup and must exit Caret.

Also, when I was playing around trying to solve John's problem, it seemed that changing the Orientation tab's X selection to right-to-left and pressing Convert to LPI didn't x-flip the volume as expected. In the past, this button has worked as far as actually flipping the volume, but the slice display didn't update. So I would then save my x-flipped volume file and reopen it in Caret, replacing any existing structural volumes that were currently loaded. This used to work. In my case just now, it didn't, and I tried using two different input volumes.

But there are a few ways to solve your problem. If Brainvoyager can export NIfTI, then the header should provide Caret the orientation information it needs to do the appropriate flipping upon opening the volume. If Brainvoyager can't export NIfTI, then complain -- loudly.

Other options include using AFNI utilities to convert Analyze to AFNI, then follow up using 3drefit to specify the existing orientation information. Then you can use 3daxialize to actually do the flipping, although I'm guessing it shouldn't be necessary (i.e., I think Caret will flip it to LPI upon opening).

Finally, try playing around with caret_command's -volume-set-orientation feature. Using FSL's very handy avg152T1_LR-marked.hdr, I found that this command caused Caret to display the right on the left:

caret_command -volume-set-orientation LPI avg152T1_LR-marked.hdr avg152T1_LR-marked_LPI+orig.HEAD

The attached right_on_right.jpg is what this volume looks like before running caret_command. Since FSL uses a left-handed coordinate system (right-on-left), they use a negative pixdim[1] to specify the x-flip:

pixdim[1]: <-2.0000>

Caret knows to x-flip the volume when it sees negative pixdim[1] in the Analyze .hdr. But if you use caret_command to tell Caret "no, the orientation is actually LPI" (as the above command is doing), then Caret won't x-flip the volume, so you get the result in the right_on_left.jpg capture. Sorry if this is confusing, but it does serve to make the point that it's important to make sure your orientation is correct, using some unambiguous anatomical landmark (or the old vitamin E capsule trick). Occipital petalia and posterior insula are common asymmetries that you can use to check orientation, if you don't have a more obvious landmark.




caret-users mailing list

caret-users mailing list

Donna L. Dierker
(Formerly Donna Hanlon; no change in marital status -- see 
http://home.att.net/~donna.hanlon for details.)

Reply via email to