I am trying the polhemus digitizer for our MEG data. The brainstorm digitizer helped generated .pos file. Then I moved the .pos file to .ds MEG data. However, when I "review raw file", before and after "refine registration", the head shape generated by digitizer did not match the MRI head model. Head shape generated by digitizer is 3-4 times bigger than MEG sensors. Images attached. What could I do?
If you don’t record EEG and MEG at the same time, you will analyze them separately.
Therefore you should have two separate .pos files and import them separately, in different folders.
Thank you so much for the suggestion.
Yes, I followed the instructions in the tutorial and the points look good there. Is there a possibility of mis-match of units? (for example, digitizer record by inches, and brainstorm sensors data are in cm.) If that is cause the problem, how could I switch the digitizer units?
Thank you,
Best,
Hi Xianghong,
It looks like the size and orientation are incorrect. After collecting the first 6 points (each coil twice), the brainstorm digitizer displays a silhouette of the default head surface with the collected points. Can you please attach a screen shot of this display?
Thank you very much. Our polhemus is in another building. I will do that later.
I have a quick question: The HPI coil location in digitizer, I heard that it is for Elekta system. Our MEG is CTF system, and do not have the HPI system. What should we do while digitize? Could that be part of the reason for the incorrect orientation?
Thank you,
Best,
The Brainstorm digitizer is for any system. We use the HPI nomenclature, but it is exactly the same as the HLC (Head Localization Coils) for the CTF system. You need to first digitize the HLC coils, then the anatomical locations, then EEG (if applicable), then head shape. You should consider reading the troubleshooting tips at the bottom of the tutorial page: http://neuroimage.usc.edu/brainstorm/Tutorials/TutDigitize
We also work primarily with CTF MEG systems.
By “HPI” we mean “head localization coils”, this refers to the three coils (blue/green/red) you stick on the subject’s head for detecting its position in the MEG helmet.
Thank you so much for the help.
I finally used a ctf software digitize, and saved .pos file. After I put the .pos file inside .ds folder, and review raw files (MEG recordings), I could see a better co-registration. The .pos file is measured from a volunteer, and the MRI and MEG is from different person, their head shape is similar. The image attached.
I digitized the head shape (we do not have simultaneous EEG recording) from brainstorm. The digitization of head shape is from one volunteer, the MEG and MRI is from another person with similar head shape. When I review the raw data (MEG recordings), the co-registration image was attached. It is still a little off.
The subject’s head looks indeed far out from the helmet, but it doesn’t look absurd, sometimes the subjects are not well positioned…
Do you have any way to check if this is or not the real position of the subject?
I suggest you run some more tests: do the full Polhemus digitization with Brainstorm AND record one minute or recordings on the same volunteer, making sure that his/her head is as adjusted to the MRI as possible.
(use a subject for whom you already have the MRI and the FreeSurfer segmentation)
Thank you so much. Just realize that I might put a different .pos file inside the .ds folder.
I saved the head shape digitization file in .elp format. Is there a way to convert the .elp file back to .pos file format?
Thank you so much,
Best,
You can write manually the .pos file with the information from the .elp file. Those are two very simple file formats…
To get an example of valid .pos file, check the folders in the CTF-based tutorial datasets on the Brainstorm website (in sample_auditory.zip or sample_raw.zip).
Hi Xianghong,
It looks like your Head localization coils and anatomical locations are exactly the same. It also looks like those points are not correctly aligned with the axis of the brainstorm coordinate system. Please follow the instructions carefully if you are using HPI coils and anatomical locations:
Collect the three coil positions (nasion, left, right) twice - these are places near the anatomical locations - not on them.
Collect the anatomical locations (nasion, LPA, RPA) twice - these are not the coil locations, but the actual anatomical locations that you will select on the MRI during the anatomical import process. See attached illustration. The red, blue and green donut shapes are the coils and the black crosshairs are the anatomical locations.
Import anatomy to brainstorm and mark the ANATOMICAL locations
Import the recording and do the automatic registration
Note: It is always a good idea to take photos of the coil placement. Therefore, if you have problems with the .pos file, you can disregard use of the .pos file and mark the coil locations directly on the anatomy.
Let me know how it goes.
Beth
Hi Xianghong,
After digitizing the head, you must be very careful when choosing the points correctly on the MRI in brainstorm. If you use the ctf software to digitize, then you must mark the locations of the COILS on the MRI (you can use photos for this). If you use the brainstorm software to digitize and you collect both the coil locations and the anatomical locations, then you mark the anatomical locations on the MRI. The difference between these two systems is that brainstorm does a transformation from the coils to the anatomy, but CTF does not.
I would also suggest you collect more points (approx 100). between the brow and the inion.
Beth
