Back to the main page.
Bug 1465 - acticap 3d electrode positions needed
| Status | CLOSED FIXED |
| Reported | 2012-05-10 13:28:00 +0200 |
| Modified | 2012-08-23 10:33:58 +0200 |
| Product: | FieldTrip |
| Component: | core |
| Version: | unspecified |
| Hardware: | PC |
| Operating System: | Windows |
| Importance: | P3 normal |
| Assigned to: | Robert Oostenveld |
| URL: | |
| Tags: | |
| Depends on: | |
| Blocks: | 1553 |
| See also: | http://bugzilla.fcdonders.nl/show_bug.cgi?id=1553 |
Johanna - 2012-05-10 13:28:21 +0200
Include 64ch and 32ch Acticap (used at DCCN) in list of 3d standard electrode positions in template/electrode, for further use with source analysis.
Robert Oostenveld - 2012-05-13 14:13:15 +0200
This has already been discussed with Madelon Vollebregt and with Sander. Sander promised to look up the documents in his archive regarding the precise layout of these caps.
Robert Oostenveld - 2012-05-13 14:13:56 +0200
Sorry, I accidentally added someone to the cc list. I removed him again.
Robert Oostenveld - 2012-05-13 14:17:23 +0200
a question to be looked into: is there any reason not to use the standard 1020 layout? Please look at the EEG1020, EEG1010 and EEG1005 and the elec1020, 1010 and 1005. I don't know why they seem to be present twice...
Johanna - 2012-05-14 08:58:44 +0200
Created attachment 265 acticap 2d layout
Johanna - 2012-05-14 08:59:39 +0200
Hi, This was not on behalf of Madelon. She is happy with the 1020 (1010 extended) layout 2D file for topoplotting. I filed this in response to a request at the toolkit from someone with EEG data wanting to do source localization, hence needed the 3D standard positions (although I recommended using Polhemus to get the actual positions per subject). The acticap does not follow the 1020 positions (I attach the Acticap layout and a second plot of the positions relative to 1020).
Johanna - 2012-05-14 09:00:09 +0200
Created attachment 266 acticap 2D relative to 1010
Johanna - 2012-06-20 13:20:18 +0200
Created attachment 278 acticap layout .mat
Johanna - 2012-06-20 13:21:39 +0200
As part of the Krakow toolkit, Irina has made http://fieldtrip.fcdonders.nl/example/preprocessing_eeg which includes the 2D Acticap layout (which does not match any 10-20 positions). Any reason why not to include this in fieldtrip/template/layout? (i've just attached it here).
Johanna - 2012-06-20 13:27:08 +0200
oops, further comment: the layout.mat and figure in the tutorial from Irina appears to match in terms of sensor position, but not sensor label. In the cap layout I got from Sander a year ago, the sensors 1-32 were a ring in the middle (see attachment to this bug of 'acticap 2d layout' whereas Irina's shows sensors 1-32 on the right side of the head. Do we still need 2 separate layout.mat files for these 2 different label setups, even though the underlying positions are the same (I believe)?
Irina - 2012-06-20 14:07:10 +0200
(In reply to comment #9) the layout from the example script - is for the ActiCap that is used in the EEG lab in the MPI (here in Nijmegen). It is the EasyCap 64-channel system. It is exactly the same system as we use in the Donders, but the arrangement of channels is not the same (!). I am not sure if we want to put it into the Template folder. I am afraid people who record their data in Donders will be confused)) By the way, I have the layout file for the Donders 64-channel cap as well. That one I can share.
Robert Oostenveld - 2012-06-21 10:05:18 +0200
we could make the layout on the basis of http://bugzilla.fcdonders.nl/attachment.cgi?id=265 and then call it "donders_customized_acticap64.mat"
Johanna - 2012-06-21 11:07:43 +0200
To clarify, which layout is 'standard' assumed by Acticap and used outside of Nijmegen, and which is a Nijmegen-specific layout? (Between either 1-32 on the right, versus 1-32 in the centre).
Robert Oostenveld - 2012-06-22 11:22:31 +0200
(In reply to comment #10) At the moment I am looking through the preprocessing_eeg example and come across the various versions of the 64 layouts For layouts in general we have already decided that we should (where possible) distribute them with a corresponding bitmap. So in this case we should have donders_customized_acticap64.mat which is the layout donders_customized_acticap64.png which is http://bugzilla.fcdonders.nl/attachment.cgi?id=265 and idem mpi_customized_acticap64.mat mpi_customized_acticap64.png with the corresponding schematic drawing. The bitmaps should (if possible) not only contain the layout, but also the comments surrounding them, i.e. the statement "actiCAP 64Ch, Customized Montage No. 10 for FC Donders Centre, 11-2007" should just be in the bitmap. That allows us to (in a few years) again identify precisely which is which.
Robert Oostenveld - 2012-06-23 11:37:09 +0200
I have added the mpi version that I received from Irina. mbp> svn commit mpi_customized_acticap64.* Adding (bin) mpi_customized_acticap64.mat Adding (bin) mpi_customized_acticap64.png Transmitting file data .. Committed revision 6118. Note that I renamed them to all lower-case letters.
Robert Oostenveld - 2012-06-23 12:11:46 +0200
I took the dccn version from Irina, and modified it manually in MATLAB (changing labels, adding the GND, etc) and then added it to the fieldtrip release. mbp> svn commit Adding (bin) layout/dccn_customized_acticap64.mat Adding (bin) layout/dccn_customized_acticap64.png Transmitting file data .. Committed revision 6119. Note that I have not yet received the final word from Sander.
Robert Oostenveld - 2012-06-23 12:15:26 +0200
(In reply to comment #15) oops, at my previous commit I included the wrong bitmap for the dccn_customized_acticap64, it was the M10 instead of the first attachment to this bug. fixed it! mbp> svn commit Sending layout/dccn_customized_acticap64.png Transmitting file data . Committed revision 6120.
Robert Oostenveld - 2012-06-23 12:25:05 +0200
Created attachment 280 screenshot of dccn and mpi version I have attached a screenshot, with in the left column the dccn_customized_acticap64 and in the right column the mpi_customized_acticap64 layout (both in MATLAB figure and in the original bitmap). From the figure it now becomes clear that the MPI version still needs some editing. The outermost channels (e.g. 26 = Oz) don't lie on the rim in MATLAB, whereas they do in the bitmap. Also REF (left mastoid) is missing, as is GND. For RM (right mastiod, for which the correct name according to the 10-20 standard from Jasper 1958 would have been M2) it would be good to include it twice, once with the label "RM" and once with the label "32". The positions are not different in the two versions (and is also the same as the M10 positions), it is just the labeling. So I'll take the dccn_customized_acticap64.mat file, copy it to mpi_customized_acticap64.mat and change the labels. ...
Robert Oostenveld - 2012-06-23 13:07:06 +0200
Created attachment 281 screenshot of mpi version I have updated the mpi_customized_acticap64.mat file with the positions from the dccn version, i.e. with the electrodes properly on the outer rim of the head (as in the bitmap). mbp> svn commit Sending layout/mpi_customized_acticap64.mat Transmitting file data . Committed revision 6121. attached is a screenshot that shows the MPI bitmap and the content of the MPI mat file.
Johanna - 2012-06-24 23:06:39 +0200
(In reply to comment #18) Hi Robert, this final version of the MPI (and the DCCN version) look fine to me, thank you. I re-assign the bug to you, so you can get the credit for resolving it. :-)
Robert Oostenveld - 2012-06-25 08:56:10 +0200
Johanna and Irina are satisfied with the result.
Robert Oostenveld - 2012-06-25 09:39:27 +0200
(In reply to comment #20) Sander is also satisfied (personal communication by email)