[Eeglablist] Choosing right head model in Dipfit with BrainProducts Cap

[Tarik S Bel-Bahar]

Greetings Maya, yes those dipoles look way too symmetric in comparison
to a range of published results. I am not saavy about the dipfit
mechanics here, but here's some notes directly below that might be
useful. Cheers!


**************************************************
It's unusual to have a problem with picking the MNI head model,it's a
simple step,  this might be a glitch. If it repeats or is not
resolved, consider adding the question to the eeglablist bugzilla.
Double check you have one clean installation of a recent eeglab in
your matlab path, and the fieldtrip plugin for eeglab is updated (via
the gui).

Consider trying the warping option in the dipfit gui if you haven't,
where you should be able to match each electrode to a specific 10-20
or 10-10 location.
Alternatively, try the warping with just the fiducials.

Alternatively you may attempt to rename each electrode to have it's
appropriate label. This can be done via channel locations gui.

Double check that you have truly loaded the correct locations, and
they show correctly in 2d and 3d channel plots. Load it again to be
sure in the channel locations gui, and check for errors.

if you haven't had a chance to, google "eeglablist channels dipfit
co-registration" for some past messages about co-registration and
dipfit, and similar topics.

Also, then, please try again to modify the channel locations by hand,
it should be easy to do by messing around for an hour or two maximum,
usually it takes much less time.
You may want to send a specific example about that specific
problem/step. They don't have to be perfect for now, just try to get
all the channels near the head surface as close as possible to their
correct location. This should lead to a good head model and good
dipfit results. You can refine the channel locations step after you've
achieved sanity. You may also want to share a picture of that if
questions remain.

Last, please remember that with 32 channels there are some sparse-EEG
caveats in interpreting source locations,even though 32-channels
source solutions get published all the time.

Overall, good luck with your adventures!


**********************************************************








On Wed, Jul 27, 2016 at 6:50 PM, Maya Pinacle <mayapinacle at gmail.com> wrote:
> Hi all,
>
> I collected my data with 32 channel BrainProducts cap, and would like to
> locate dipoles via Dipfit. I tried to use manual Co-registration option
> choosing BEM MNI as head model however unfortunately it is not working well
> and it is very difficult to align the electrode template with my own
> electrode locations.
>
> However tutorial says that if the electrode locations within the
> Internaitional 10-20 system best way is to use standard channel coordinates
> associated with the head model. So as I know BrainProducts caps are in 10-20
> systems so I used this option with recomended settings, choosing Boundary
> Element Mode (MNI) as Head model with No Coregistration to be chosen.
>
>
>  The tutorial also says to go to channel locations from edit option of
> EEGLAB and look up for the channel locations. In which case I choose ''use
> MNI coordinate file for BEM dipfit model'' option. In each condition (Coarse
> fine or auto fit ) the coordinates of the dipoles seems sort of overlapping
> rather than spreaded all over the brain. And when I create STUDY from
> different subjects file and cluster the component dipoles they lined up in a
> very straight coordinates (I have added 4 pictures.)
>
> When I compare visually a component and a dipole that represents it, it
> seems that estimation of the dipole location is pretty good with those
> selection of options as it was described in the tutorial. However, this very
> well organised appearance of the dipoles in a cluster made me think that
> whether something wrong with it. Because it seems like rather than finding
> actual source, the dipoles try to locate around the channels and I wonder
> whether the result is kind of misleading. So I wonder is there anyone would
> like to comment on that? I can also share the data in the case you would
> like to see.
>
> So from one data the clusters and dipole locations can be found here.
>
> Cluster 1 - 33 dipoles - Sagital
> https://www.dropbox.com/s/rf77qcznrk7bvia/cluster1_sagital.png?dl=0
>
> Cluster 1 - 33 dipole - Top
>
> https://www.dropbox.com/s/9l2p7l7kmlsz7r0/cluster1_top.png?dl=0
>
> Cluster 2 - 62 dipole Sagital
>
> https://www.dropbox.com/s/cphtiemq05xeacu/cluster2_sagital.png?dl=0
>
>
> Thank you
>
> Maya
>
> Cluser 2 - 62 dipole Top
> https://www.dropbox.com/s/64ofdpixqo6bw4k/cluster2_top.png?dl=0
>
> _______________________________________________
> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to eeglablist-unsubscribe at sccn.ucsd.edu
> For digest mode, send an email with the subject "set digest mime" to
> eeglablist-request at sccn.ucsd.edu