Dear Laura,

There is no requirement that the electrodes should be in the same coordinate system as the MRI to begin with but the fiducials should be specified in the coordinate system of your individual image in SPM. I would suggest that you do include fiducials with your electrode locations and then open your nifti structural in SPM (Display/image), find the corresponding points in the image, write down their coordinates and use that for coregistration. The coordinates you have from ASA software will not work. There are more sophisticated ways to transfer your coregistration to SPM but if you can find your fiducials in the image reliably enough, that'll work as well.

Best,

Vladimir

On Thu, Jul 3, 2014 at 11:08 AM, Laura Rueda Delgado <[log in to unmask]> wrote:

Dear SPM users,

I’m working on source estimation in EEG (128 channels). For creating the headmodel, I have the individual MRI and the individual electrode locations. I defined manually the fiducials (nasion, LPA, RPA) in the MRI images using the ASA software (from ANT). These fiducials and the electrode positions are in the CTF MRI coordinate system (x-anterior, y-left, z-superior). The electrode positions were measured with respect to the same fiducials, but they’re not included in the electrode file, so its size is of 128 points. When registered with respect to the fiducials taken from the MRI, the electrodes appear where they should in fieldtrip. However, I’d like to benefit from SPM’s MRI segmentation and mesh creation algorithms.

When I load the MRI images in SPM, I transform them from DICOM to NIFTI using the SPM importer. The electrode positions and the EEG data are read first via fieldtrip, and then I use spm_eeg_ft2spm(). Then I load the sensors and fiducials via the corresponding functions. Then I segment and generate the meshes using spm_eeg_inv_mesh(). And then I use spm_eeg_inv_datareg() to coregister the location to the meshes.

There are no errors when running the script, but the coregistration does not work properly. It is very clear that the electrodes are oriented differently from the scalp mesh, as you can see below.

I think the problem is that the electrode and fiducials location need to be transformed to the corresponding NIFTI (or SPM) coordinate system, which is the coordinate system of the mesh. However, I’m not sure if that’s the case or if so, how I could fix it. Anybody has any ideas?

Thank you in advance for your help.

Best regards,

Laura Rueda