beamformer source localization for EEG
am236 at GEORGETOWN.EDU
Thu Jul 9 04:43:39 CEST 2009
Dear Fieldtrip developers/users
I am a relatively recent user of Fieldtrip (and very much appreciative of this software). Having good experience with the analysis of brain oscillations etc, I now want to proceed with source reconstruction and it appeared that I need some help with the beamformer technique for EEG data. I am trying to use it for EEG data collected with a 128-channel EGI system. I also have structural MRI data for my subjects and would like to use realistic head models etc. I was following the Fieldtrip tutorial (1) “Localizing oscillatory sources using beamformer techniques”, which applies this method to MEG data. My understanding is that the only difference I would make when applying this approach to EEG is that, instead of the command
vol = prepare_singleshell(cfg, segmentedmriF);
which implements Nolte’s approach for MEG data, I need to make a BEM model for EEG data and I thought that I could use an example Matlab script (2) “Create BEM headmodel for EEG” from the Fieldtrip site. This would give me a vol variable which I then use to calculate the grid variable using
[grid] = prepare_leadfield(cfg);
and the rest is the same as in tutorial (1).
Q1. Is it a correct sequence of steps to do source reconstruction for EEG data in Fieldtrip or do I need to do something more different?
Q2. My second question is about the function strel_bol() which I am not familiar with (it is used in script (2)). One can see from the script that it is a structural element function used in the Image Processing toolbox but this toolbox has only function strel(). Could you please explain what the strel_bol(5) is? Can it be replaced by something like strel('disk',5)?
Your help is very much appreciated!
Andrei Medvedev, PhD
EEG and Optical Imaging Lab
Center for Functional and Molecular Imaging
Washington DC, 20057
More information about the fieldtrip