Head Localization and MRI Coregistration: Difference between revisions

From MEG Core
Jump to navigation Jump to search
Content added Content deleted
No edit summary
No edit summary
 
(6 intermediate revisions by the same user not shown)
Line 3: Line 3:
The first step in any analysis pipeline is to localize the MEG data to anatomical space, which is generally done using a structural MRI. When doing this localization, it is important to know if the subject moved during the recording.
The first step in any analysis pipeline is to localize the MEG data to anatomical space, which is generally done using a structural MRI. When doing this localization, it is important to know if the subject moved during the recording.


* Use headMotionDetect to determine epochs with head motion over a given threshold (in cm)
* Use [[local:/meg/headMotionDetect|headMotionDetect]] to determine epochs with head motion over a given threshold (in cm)


{| class=wikitable width=600 align=center
: {| width=600 style="border: 1px solid lightgrey;"
|headMotionDetect -th 0.25 dataset.ds
|Usage: headMotionDetect -th 0.25 dataset.ds
|}
|}


* The first step in coregistration of MRI and MEG data is to locate the three fiducial points on an MRI scan. This can be done in one of two ways:
*
#Manual Method: Manually locate the positions of the three fiducial coils. This can be facilitated by using MRI-visible fiducials at the time of the scan, or vitamin E capsules taped to the head at the fiducial points. Use the Edit Tagset plugin within the AFNI viewer to add "Nasion," "Left Ear," and "Right Ear" tags.
#Digitizer Method: At the time of the MEG, digitize the location of the head coils on the subject's head, along with either additional digitizer points covering the head or a 3d scan of the head shape (with a Polhemus device, Kinect, or Structure sensor). The headshape point cloud can then be aligned with the MRI in a semi-automated fashion; a detailed tutorial for this method is forthcoming.

* The tool [[orthohull.py|orthohull.py]] is used to create the headmodel needed for further processing from the AFNI anatomical MRI with fiducial points "Nation," "Left Ear," and "Right Ear" saved as tags in the header. It is assumed that the MRI has not been skull-stripped. The -t and -m options perform a normalization to Talairach or MNI space, respectively. Additional options available, execute command with no arguments to see usage information.
: {| width=600 style="border: 1px solid lightgrey;"
|Minimal usage: orthohull.py anat+orig.HEAD
|}
: You can additionally use SUMA to visualize the output surface, which defaults to ortho_brainhull.ply
:{|width=600 style="border: 1px solid lightgrey;"
|suma -i_ply ortho_brainhull.ply
|}
: You can also visualize the overlay of the surface on the anatomical; once SUMA starts, press "t" to talk to AFNI.
:{|width=600 style="border: 1px solid lightgrey;"
|afni -niml -dset ortho+orig mask+orig
|suma -niml -i_ply ortho_brainhull.ply -sv ortho+orig -novolreg
|}

* The CTF and SAM pipelines can use either a Multisphere headmodel, or a Nolte (realistic head shape) headmodel. Both require a "hull.shape" file, the multisphere model requires additional .shape and .shape_info files (in a different format from hull.shape). The orthohull tool will create the hull.shape file (using [[local:/meg/meshnorm|meshnorm]]) and the .shape and .shape_info files (using [[local:/meg/ply2fid|ply2fid]).

* [[local:/meg/localSpheres|localSpheres]] is used to compute the final required files for the multisphere headmodel; this file is specific to the MEG dataset. Use the -help option to see additional details, sometimes the -r option is necessary.
:{|width=600 style="border: 1px solid lightgrey;"
|localSpheres -d dataset.ds filename.shape -v
|}
: You can check if the spheres are sane using [[local:/meg/checkSpheres|checkSpheres]]
:{|width=600 style="border: 1px solid lightgrey;"
|checkSpheres dataset.ds
|}

Latest revision as of 15:23, 19 March 2019

Head Localization and MRI Co-Registration

The first step in any analysis pipeline is to localize the MEG data to anatomical space, which is generally done using a structural MRI. When doing this localization, it is important to know if the subject moved during the recording.

  • Use headMotionDetect to determine epochs with head motion over a given threshold (in cm)
Usage: headMotionDetect -th 0.25 dataset.ds
  • The first step in coregistration of MRI and MEG data is to locate the three fiducial points on an MRI scan. This can be done in one of two ways:
  1. Manual Method: Manually locate the positions of the three fiducial coils. This can be facilitated by using MRI-visible fiducials at the time of the scan, or vitamin E capsules taped to the head at the fiducial points. Use the Edit Tagset plugin within the AFNI viewer to add "Nasion," "Left Ear," and "Right Ear" tags.
  2. Digitizer Method: At the time of the MEG, digitize the location of the head coils on the subject's head, along with either additional digitizer points covering the head or a 3d scan of the head shape (with a Polhemus device, Kinect, or Structure sensor). The headshape point cloud can then be aligned with the MRI in a semi-automated fashion; a detailed tutorial for this method is forthcoming.
  • The tool orthohull.py is used to create the headmodel needed for further processing from the AFNI anatomical MRI with fiducial points "Nation," "Left Ear," and "Right Ear" saved as tags in the header. It is assumed that the MRI has not been skull-stripped. The -t and -m options perform a normalization to Talairach or MNI space, respectively. Additional options available, execute command with no arguments to see usage information.
Minimal usage: orthohull.py anat+orig.HEAD
You can additionally use SUMA to visualize the output surface, which defaults to ortho_brainhull.ply
suma -i_ply ortho_brainhull.ply
You can also visualize the overlay of the surface on the anatomical; once SUMA starts, press "t" to talk to AFNI.
afni -niml -dset ortho+orig mask+orig suma -niml -i_ply ortho_brainhull.ply -sv ortho+orig -novolreg
  • The CTF and SAM pipelines can use either a Multisphere headmodel, or a Nolte (realistic head shape) headmodel. Both require a "hull.shape" file, the multisphere model requires additional .shape and .shape_info files (in a different format from hull.shape). The orthohull tool will create the hull.shape file (using meshnorm) and the .shape and .shape_info files (using [[local:/meg/ply2fid|ply2fid]).
  • localSpheres is used to compute the final required files for the multisphere headmodel; this file is specific to the MEG dataset. Use the -help option to see additional details, sometimes the -r option is necessary.
localSpheres -d dataset.ds filename.shape -v
You can check if the spheres are sane using checkSpheres
checkSpheres dataset.ds