Example on SPM Faces dataset - 2 questions

✉️ Mailing List Archive (read-only)
1

Dear MNE users,

I came back with a problem using mne.gui.coregistration() with CTF data.

I loaded the spm sbj and the raw file
SPM_CTF_MEG_example_faces1_3D_raw.fif and

1) I couldn't see Fiducial points
2) I couldn't see the MEG sensors

About point 1) the problem is that Christian Brodbeck wrote in the
email I attached.
If I change the 'kind' key of raw.info['dig'] to FIFF.FIFFV_POINT_CARDINAL
I'm able to see the fiducial points in the GUI, even if I think there
is probably
a mistake in assignment the nasion, lpa and rpa.
I think the right assignment should be

raw.info['dig'][0]['ident'] = 2
raw.info['dig'][1]['ident'] = 1
raw.info['dig'][2]['ident'] = 3

Does it make sense?

About point 2) I thought the problem could be the 'coil_type' of
raw.info['chs'] has
the value 201609. I tried to change this value but it didn't happen when
I import the data in the GUI.

I also tried to perform the co-registration in mne_analyze but when I
import the
raw file I get the following error message
No suitable coordinate transformation found in
SPM_CTF_MEG_example_faces1_3D_raw.fif

Thanks for any suggestions!

best regards
Annalisa

----- Messaggio inoltrato da Christian Brodbeck
<christianbrodbeck at nyu.edu> -----
Rispondi-A: Discussion and support forum for the users of MNE Software
<mne_analysis at nmr.mgh.harvard.edu>

2

Hi Annalisa,

the landmarks are transformed to RAS coordinate system when the data is
converted to fif format. As the constant definitions map to nasion = 2, lpa
=1 and rpa = 3 ('ident' in chs) they seem correct to me.

The second point seems like a bug to me. I've opened an issue here
https://github.com/mne-tools/mne-python/issues/3071

-Jaakko

3

Hi Jakko,

thks a lot for your answers.

About point 1), when I read the fif file, I get this info

raw.info['dig']
Out[89]:
[{'coord_frame': 4,
   'ident': 3,
   'kind': 2,
   'r': array([ 4.30736691e-09, 9.82968360e-02, 0.00000000e+00],
dtype=float32)},
  {'coord_frame': 4,
   'ident': 1,
   'kind': 2,
   'r': array([ -7.32455477e-02, -3.20142135e-09, 0.00000000e+00],
dtype=float32)},
  {'coord_frame': 4,
   'ident': 2,
   'kind': 2,
   'r': array([ 7.07898811e-02, 2.96859071e-09, 0.00000000e+00],
dtype=float32)}]

I think there are 2 bugs, since

i) the value of 'kind' it would be FIFF.FIFFV_POINT_CARDINAL=1
    while now it is FIFF.FIFFV_POINT_HPI=2
    Indeed, I cannot see these points in the GUI if I don't change these values
    and write the raw file with this update information

ii) Furthermore also ident values don't seem correct...the right value
would be
     raw.info['dig'][0]['ident'] = 2
     raw.info['dig'][1]['ident'] = 1
     raw.info['dig'][2]['ident'] = 3

Are u agree?

About my second point I'll follow the issue

Thks a lot!

kind regards,
Annalisa

Jaakko Lepp?kangas <jaeilepp at student.jyu.fi> ha scritto:

4

Hi Annalisa,

In mne_analyze, you

1) Load the FreeSurfer surfaces (Load surface?)
2) Load the digitization data from the file rather than try to open the raw file (Load digitizer data?)

The rest of the alignment steps are described in Section 12.11 of the MNE manual.

Best,
Matti

5

Dear Matti,

thanks for your suggestion.

I tried to do it but I see only three green circles.

The file I used is SPM_CTF_MEG_example_faces1_3D_raw.fif

What's wrong?

Thanks for hep me

Best regards,
Annalisa

Matti Hamalainen <msh at nmr.mgh.harvard.edu> ha scritto:

6

Annalisa,

The green circles represent the fiducials, so they are showing up
properly in mne_analyze.
What are you expecting to see?

hth
d

7

hi Denis,

yes, you are right!
I can see only these 3 points since
there are not other digitization points.
Thus, I could perform the co-registration only using
the fiducials.

I waited to see also the helmet importing the raw data
but I take this message error
"No suitable coordinate transformation found in
SPM_CTF_MEG_example_faces1_3D_raw.fif"

Thanks for your answer!

Best,
Annalisa

The green circles are identified
dgw <dgwakeman at gmail.com> ha scritto:

8

FYI

we provide in the spm dataset a trans file you can use. You can
rely on it and avoid the coregistration.

File is called:

MNE-spm-face/MEG/spm/SPM_CTF_MEG_example_faces1_3D_raw-trans.fif

HTH
Alex

9

Hi Alex,

yes, I have seen this file, but I asked you since
I need to perform the co-registration for other
CTF files that I converted in .fif and thus I
wanted to understand the correct procedure.

So, since mne.gui.coregistration doens't work correctly
I'll use mne_analyze with only the 3 fiducial points.

Thanks!
best,
Annalisa

Alexandre Gramfort <alexandre.gramfort at telecom-paristech.fr> ha scritto: