[FieldTrip] Head segmentation

[RICHARDS, JOHN]

Jeff

The resources from Carsten Wolters and the answers were right on.

We have a 128 channel system.

We add “nasal cavity” which includes nasal, throat, and sinuses in our FEM model.  We do a lot of the segmenting with FSL (bet, fast, betsur, SPMf) and custom programs, and then do manual editing to identify the air passages.  The separate segments are then put together with a custom program. The air cavities might be important for the 256-channel EGI system, where on the sides and front of the face the electrodes could have some dipolar sources and the places in between would affect the source estimation.  Anything below the skull, not air cavity, we give a conductivity of muscle, which is very close to scalp; again the electrodes below the skull are over something and I presume would need a realistic estimated head media and conductivity.

As I wrote earlier, we have been using FT for FEM recently; also have gotten good (similar to FEM) results with the Dipoli BEM model. I have one already-published supplemental information that briefly describes our FEM model, and one in-press supplemental information, both articles using the EMSE program.  I am working on our first publication using the FEM from FT.

Re “Mainly tissues between sources and electrodes are important to be modeled (so white matter anisotropy is only important for deeper sources)”.  This may be practically correct, but I have suspected (and EGI’s rationale for whole-head coverage) that to get a good measure of a dipolar source whose positive or negative pole land somewhere below the skull area, one needs a good model of the head over all the places where the current flow is manifested.  We have some prefrontal  sources with large negative potentials exhibiting on the frontal-central 10-10 electrodes, and which have a large positive scalp potential on the “Oz” electrodes.  However, it turns out these potentials show up on Iz and in the EGI electrodes below Iz (128 channel), so having a model for scalp and muscle over the neck might be important.  I don’t know what the effect is of leaving these out, but I have preferred to do the whole head segmentation in case these have some effect.

One last consideration.  Most BEM and FEM models on the FT site use the segmentation of the brain, skull, and scalp about at the level of the MNI head.  This does not include the electrodes below the horizon of the MNI head, and there is a lot of area below the MNI head that is covered by the EGI 128 and 256 channel systems.  This adds significantly to the volume of the voxels used for the FEM estimation.  For BEM models we cut the head off at about 10% below the level of the lowest EGI 128 channel electrodes, and generally have done so also for the FEM models we use in FT. The electrodes are located on the places below the MNI head, so one has to have a “head” model with positions below those typically used in the FT’s demo programs.  We do a whole head MRI down to the neck, and then use a custom program to get a MRI volume with a head that extends about 10% lower than the lowest EGI-128 electrodes.

John


***********************************************
John E. Richards Carolina Distinguished Professor
Department of Psychology
University of South Carolina
Columbia, SC  29208
Dept Phone: 803 777 2079
Fax: 803 777 9558
Email: richards-john at sc.edu<mailto:richards-john at sc.edu>
HTTP: jerlab.psych.sc.edu
***********************************************

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20151202/8ecc7956/attachment-0001.html>