Dear Vivek,

On Fri, Oct 7, 2011 at 3:15 AM, Buch, Vivek (NIH/NINDS) [G]
<[log in to unmask]> wrote:
> Hi Dr. Litvak,
> Thank you so much for the reply. I did indeed find this out because I realized that I can define my own nifti image as the sMRI input in the spm_eeg_inv_mesh function. However, in this function I am not asked to define the MEG fiducials or define the voxel-space for the T1 fiducials, which I somehow need to do.
>

See below

> I understand what you are saying about coregistering the subject T1 to the SPM template. But I feel that for our purposes I might >prefer to coregister the subject MEG to the subject T1 directly if possible, so that we get the best anatomical representation of >source space for the DCM.

Coregistering an image to the template will not affect the 'anatomical
representation of the source space'. It is not the same as
normalisation. If your images are slightly off it will be OK, but if
they are, for instance rotated with respect to the template, you won't
be able to use them in SPM so this is something you must check and fix
if necessary.


> I just failed to mention in the previous email, but we do get the same 3 fiducials coordinates in MEG space via CTF, as a .res4 file > within our .ds file.
>

Not sure what you mean, but it's not enough if can just read the
fiducial coordinates off the screen. They must be in the dataset in
the particular representation that we use so you should follow the
instructions from the previous e-mail.


> Ideally, I would like to write a script to be able to batch the coregistration of each subject's MEG sensor space to their own T1 image over all of our subjects, and would therefore actually prefer to use the matlab functions rather than the GUI. I know this would be possible to do, but am just having difficulty with the details! I think I may need to use a combination of the spm_eeg_prep and spm_eeg_inv_mesh functions--but cannot figure out exactly how to define the MEG.res4 fiducial file, T1 voxel coordinates, and T1 nifti image within those matlab functions to be able to coregister the subject MEG space to the subject T1 space to create an accurate head model for each subject. Would you be able to recommend a course of action?
>

You should look at the SPM batch tool. There is a tool for 'head-model
specification' there. Once you make it work for your data you can do
'Save as...' and choose '.m' file. This will generate the
corresponding Matlab code that you can put in your script, add the
line

spm_jobman('run', matlabbatch);

and you are ready to go.

> Thank you so much, I am very sorry to bother you on this subject.

You are not bothering me at all. It's part of my job to answer
questions. However, I'd prefer that you send questions of general
interest to the SPM mailing list ([log in to unmask]).

Best,

Vladimir

>
> Best,
> Vivek
>
> ________________________________________
> From: Vladimir Litvak [[log in to unmask]]
> Sent: Thursday, October 06, 2011 7:10 PM
> To: Buch, Vivek (NIH/NINDS) [G]
> Cc: [log in to unmask]
> Subject: Re: Question regarding coregistration in SPM8
>
> Dear Vivek,
>
> What you are overlooking is that nifti format is the native format of SPM so there is no need to convert it at all. What you should make sure is that your images are roughly aligned to SPM template or it might be a good idea to coregister and reslice them to the template using functionality in fMRI part of SPM. Then you can just press the MRI button and select your coregistered image.
>
> There is a different issue that might trouble you. You need fiducials for SPM coregistration and these are not present in Fieldtrip data. You can look at example script in man/example_scripts to see how you can add them, or there is also a tool in MEEGtools toolbox for that (Copy MEG sensors).
>
> Best,
>
> Vladimir
>
>
>
> On 6 Oct 2011, at 21:59, "Buch, Vivek (NIH/NINDS) [G]" <[log in to unmask]<mailto:[log in to unmask]>> wrote:
>
> Hi Dr. Litvak,
> My name is Vivek and I am a new research fellow at the NIH conducting cortical plasticity research using MEG. I am very interested in using DCM to start modeling effective connectivity graphs in my data. However, I am having trouble with the preparation. I have preprocessed the MEG data in fieldtrip (we collect MEG using CTF) and converted to spm format. However, I am having trouble with the coregistration process. I have standard NA, LPA, and RPA fiducial coordinates in voxel space for each subject’s T1 scan. The scan format is nifti (.nii). I need to coregister the MEG sensor space to the nifti T1 space for each subject, but I can’t figure out how to get the .nii format converted to either .mri or some other format that is readable by SPM. I have tried looking into fieldtrip tools but their ft_volumerealign does not recognize .nii files. Is there something I am overlooking, or are there other ways for me to go about doing this?
>
> Very sorry to bother you, thank you very much for your help.
>
> Best,
> Vivek
>
> Vivek P. Buch, MD Candidate
> HHMI-NIH Research Fellow
> National Institutes of Health, Bethesda, MD
> Warren Alpert Medical School of Brown University, Providence, RI
>