[pymvpa] How to export neighborhood relationships of NiftiDataset?

Thu Nov 10 10:36:40 UTC 2011

Dear Thorsten,

         Thank you for your reply. What I have done may be something curious. I did slice timing, realignment, coregistering, and normalization using SPM. Then I changed the .img/.hdr pairs (supplied by SPM) into 4D .nii.gz files which are more suitable for PyMVPA. For SPM, I can check the MNI coordinates easily. But for PyMVPA, I don’t know how to check the MNI coordinates. What I got using dataset.mapper.getInId() is a coordinate in the voxel space, I think. I wonder how to change them into the MNI space.

Best wishes!

     Lu FENG

fenglu at fingerpass.net.cn

From: pkg-exppsy-pymvpa-bounces+fenglu=fingerpass.net.cn at lists.alioth.debian.org [mailto:pkg-exppsy-pymvpa-bounces+fenglu=fingerpass.net.cn at lists.alioth.debian.org] On Behalf Of Thorsten Kranz
Sent: 2011年11月10日 18:06
To: Development and support pf PyMVPA
Subject: Re: [pymvpa] How to export neighborhood relationships of NiftiDataset?

Hi Lu Feng,

getting to some voxel-indices to MNI coordinates is a somewhat tricky task. What you have to do is (assuming you have a EPI-sequence, not structural MRI):

1) coregister your EPI to a structural image of the same subject
2) normalize the structural image to some standard image, e.g. MNI152T1_2mm as contained in fsl data
3) apply the concatenated transform EPI->structural an structural->MNI to the EPI sequence
4) thus you obtain an image that is in "MNI space", e.g. by opening it in fslview you can find out the MNI coordinates.

Going to Talairach would be the next step, look at

http://imaging.mrc-cbu.cam.ac.uk/imaging/MniTalairach  

for more details.

But be careful: Talairach is not a 1:1 analogy to MNI, some parts of the Tal-Brain are much smaller than in MNI. MNI is an average of 152 brains, Talairach is based on only one example brain.

You do not have to do steps 3+4, after obtaining the two transforms from EPI to structural to MNI, you can also manually calculate MNI coordinates from your voxel coordinates, but I don'T know if you want to do this.

Greetings,

Thorsten

2011/11/10 冯璐 <fenglu at fingerpass.net.cn>

... I can't understand your words exactly ... I used SPM before and I didn't have to think about the voxel space and MNI space, etc. Let's move to some practical problems.

Loading from
        Task1.nii.gz,
I can get
        dataset.samples which is a M*N*P matrix.
Using
        dataset.mapper.getInId(index),
I can get
        a coordinate [z, y, x], x belongs to (0, M), y belongs to (0, N), z belongs to (0, P).
I have checked the result above.

Using
       fslhd Task1.nii.gz,
I can get some information may be useful
       qform_name     Aligned Anat
               qform_code     2
               qto_xyz:1      -3.500000  0.000000  -0.000000  76.000000
               qto_xyz:2      0.000000  3.500000  -0.000000  -112.000000
               qto_xyz:3      0.000000  0.000000  4.000000  -50.000000
               qto_xyz:4      0.000000  0.000000  0.000000  1.000000
               qform_xorient  Right-to-Left
               qform_yorient  Posterior-to-Anterior
               qform_zorient  Inferior-to-Superior
               sform_name     Aligned Anat
               sform_code     2
               sto_xyz:1      -3.500000  0.000000  0.000000  76.000000
               sto_xyz:2      0.000000  3.500000  0.000000  -112.000000
               sto_xyz:3      0.000000  0.000000  4.000000  -50.000000
               sto_xyz:4      0.000000  0.000000  0.000000  1.000000
               sform_xorient  Right-to-Left
               sform_yorient  Posterior-to-Anterior
               sform_zorient  Inferior-to-Superior

Then, how can I change the coordinate into a MNI coordinate or a Talairach coordinate?

Sorry for interruption and thanks in advance.

Best wishes!

    Lu FENG
fenglu at fingerpass.net.cn

-----Original Message-----
From: pkg-exppsy-pymvpa-bounces+fenglu=fingerpass.net.cn at lists.alioth.debian.org [mailto:pkg-exppsy-pymvpa-bounces+fenglu <mailto:pkg-exppsy-pymvpa-bounces%2Bfenglu> =fingerpass.net.cn at lists.alioth.debian.org] On Behalf Of Yaroslav Halchenko

Sent: 2011年11月8日 20:55
To: Development and support pf PyMVPA
Subject: Re: [pymvpa] How to export neighborhood relationships of NiftiDataset?

axes and direction depend on how they are stored in your specific
Nifti -- and those could vary since we do not deduce physical
orientation of the volume anyhow from qform.

On Tue, 08 Nov 2011, 冯璐 wrote:

> I think that is what I need.
> So array([0, 0, 1]) is the coordinate in the voxel space. X from 0 to max: right->left; Y from 0 to max: posterior->anterior; Z from 0 to max: inferior->superior ??

> Best wishes!

>      Lu FENG
> fenglu at fingerpass.net.cn

> -----Original Message-----
> From: pkg-exppsy-pymvpa-bounces+fenglu=fingerpass.net.cn at lists.alioth.debian.org [mailto:pkg-exppsy-pymvpa-bounces+fenglu <mailto:pkg-exppsy-pymvpa-bounces%2Bfenglu> =fingerpass.net.cn at lists.alioth.debian.org] On Behalf Of Yaroslav Halchenko
> Sent: 2011年11月8日 13:35
> To: Development and support pf PyMVPA
> Subject: Re: [pymvpa] How to export neighborhood relationships of NiftiDataset?

> so we are talking about 0.4.x series of pymvpa (in 0.6 AKA mvpa2 voxel
> indices are explicitly stored in feature attributes of a dataset):  so I
> guess you are looking for:

> In [10]: from mvpa.tests.tests_warehouse import datasets
> In [11]: datasets['3dsmall'].mapper.getInId(0)
> Out[11]: array([0, 0, 1])

> ?

> On Tue, 08 Nov 2011, 冯璐 wrote:

> >    Dear experts,

> >             I  know that  using NiftiDataset.map2Nifti()  or other  functions
> >    like mapForward()  or mapReverse(),  we  can deal  with  either 2D  or  3D
> >    dataset freely. But I wonder how to export the neighborhood  relationships
> >    of the volume.  Briefly, I��d like  to have a  transformation matrix or  a
> >    list of  coordinations  for each  voxel.  I need  this  if I  go  on  with
> >    processing by other tools instead of PyMVPA.

> >             Thanks in advanceJ

> >    Best wishes!

> >         Lu FENG

> >    fenglu at fingerpass.net.cn

> >    __________ Information from ESET NOD32 Antivirus, version of virus
> >    signature database 6532 (20111010) __________

> >    The message was checked by ESET NOD32 Antivirus.

> >    [1]http://www.eset.com

> > References

> >    Visible links
> >    1. http://www.eset.com/

> > _______________________________________________
> > Pkg-ExpPsy-PyMVPA mailing list
> > Pkg-ExpPsy-PyMVPA at lists.alioth.debian.org
> > http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-exppsy-pymvpa
--
=------------------------------------------------------------------=
Keep in touch                                     www.onerussian.com
Yaroslav Halchenko                 www.ohloh.net/accounts/yarikoptic

_______________________________________________
Pkg-ExpPsy-PyMVPA mailing list
Pkg-ExpPsy-PyMVPA at lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-exppsy-pymvpa

__________ Information from ESET NOD32 Antivirus, version of virus signature database 6616 (20111109) __________

The message was checked by ESET NOD32 Antivirus.

http://www.eset.com

_______________________________________________
Pkg-ExpPsy-PyMVPA mailing list
Pkg-ExpPsy-PyMVPA at lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-exppsy-pymvpa

__________ Information from ESET NOD32 Antivirus, version of virus signature database 6617 (20111110) __________

The message was checked by ESET NOD32 Antivirus.

http://www.eset.com

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/pkg-exppsy-pymvpa/attachments/20111110/83ca7c7a/attachment-0001.html>