Nifti with Anatomist

Questions about Anatomist manipulation

Moderators: denghien, riviere

Post Reply
mdojat
Posts: 8
Joined: Tue Jun 14, 2005 4:01 pm
Location: Grenoble Institut Neurosciences
Contact:

Nifti with Anatomist

Post by mdojat »

We have troubles in openning Nifti files.

We have a structural image and fMRI images in nifti format (3D or 4D). In their respective hdr, the orientation parameters (equivalent to .mat files in SPM2) are different. With SPM5 and MRICron the images can be superimposed. With anatomist apparently the orientation parameters are not read correctly. The 4D format is not open correctly. Is there any way to solve this ?
Images are available for problem illustration.

Michel Dojat & Pierre Lafaye de Micheaux
User avatar
riviere
Site Admin
Posts: 1361
Joined: Tue Jan 06, 2004 12:21 pm
Location: CEA NeuroSpin, Saint Aubin, France
Contact:

Re: Nifti with Anatomist

Post by riviere »

Hi,

Well, NIFTI transformations are read, but not applied by default (mainly for historical backward-compatibility considerations). You can make them used by right-clicking on loaded objects in Anatomist main window and selecting "referential -> apply builtin referential (SPM/NIFTI)". You can also change your setting to make them automatically used at load-time: in the preferences, see the "Volumes" tab. This option may become the default one day...
There was a... ahem... an outdated french doc here on this topic: it has not been updated since 3.0 version when this option did only deal with SPM origin, but this has been extended to NIFTI (and MINC) transformations.

Denis
mdojat
Posts: 8
Joined: Tue Jun 14, 2005 4:01 pm
Location: Grenoble Institut Neurosciences
Contact:

Re: Nifti with Anatomist

Post by mdojat »

Hum,

We tried your suggestion: the use of "Apply building referential" but without succeeding.
We add to this email two examples of the superpostion of two images, anatomical and fonctional, using Anatomist and Mricron. As you can see the superposition is correct with the later but not the former.
Best
M. Dojat a P. Lafaye de Micheaux
superposition with mricron
superposition with mricron
mricron.jpg (157.32 KiB) Viewed 6084 times
superposition with anatomist
superposition with anatomist
anatomist.jpg (69.86 KiB) Viewed 6085 times
User avatar
riviere
Site Admin
Posts: 1361
Joined: Tue Jan 06, 2004 12:21 pm
Location: CEA NeuroSpin, Saint Aubin, France
Contact:

Re: Nifti with Anatomist

Post by riviere »

Hi,
Could you please send me both images so that I can test it ? You can use the anonymous ftp server of CEA, if you don't know the URL, tell me, I will send it to you via email or private message.
Denis
Dominique Geffroy
Site Admin
Posts: 161
Joined: Thu Mar 01, 2007 4:22 pm
Location: IFR 49 - Neurospin, Gif-sur-Yvette, France
Contact:

Re: Nifti with Anatomist

Post by Dominique Geffroy »

Hi,

maybe you have to change the referential of the window ? Indeed, using Anatomist "apply builtin referential" feature, the referentials and transformations of the header are loaded and a new referential is associated to the image. But by default, it is not the referential of the window and maybe there is no transformation between the referential of the window and the one of the image.
Could you send a screen capture of the referential window ( menu Settings -> referentials window ) ?

To see the images correctly superimposed, you have to visualize them in a common referential...

Dominique
User avatar
riviere
Site Admin
Posts: 1361
Joined: Tue Jan 06, 2004 12:21 pm
Location: CEA NeuroSpin, Saint Aubin, France
Contact:

Re: Nifti with Anatomist

Post by riviere »

Hi,
Ok I have seen your images.
Both have a NIFTI transformation to a "scanner-based anatomical coordinates" referential. It's just that in Anatomist we don't automatically assume this scanner-based referential is the same on all images because we often mix images from different subjects acquired on different machines. So here you have to manually tell Anatomist that the two scanner-based referentials are the same:
open the referentials window. Both images have its own referential (here blue for the anatomy and white for the functional image), and both have a transformation to another "scanner-based" referential (yellow and cyan). Drag a line between the yellow and cyan referentials while maintaining <ctrl> pressed: this will make an identity transformation between both referentials, meaning they are the same. Then you can visualize the volumes (or the fusion) in a window which has a referential with a transformation to any of these referentials (blue, white, yellow or cyan).
snapshot_anatomist_forum1251.jpg
snapshot_anatomist_forum1251.jpg (149.38 KiB) Viewed 6051 times
This behaviour is a little bit more complex than what most image viewers do because we want to be more generic and allow to view many volumes in Anatomist, not necessarily just 2 volumes with a common scanner-based referential, which is a common but special case...
When the NIFTI transformation leads to a really unique referential (such as the MNI normalized referential), you don't need to manually change anything in Anatomist.
Denis
Post Reply