<div dir="ltr">Dear Fieldtrip users and developers,<div><br></div><div>I've been struggling quite some time now with the following problem.</div><div><br></div><div style>We want to do source localization of MEG data from an experiment with 10 subjects. We collected MRIs using a Phillips scanner (UvA), and MEG data using the Neuromag Elekta scanner (VUmc).</div>
<div><br></div><div>Using Neuromag software in Linux (seglab, xfit), I've created BEM forward models based on coregistered MRIs (coregistration also done using the Neuromag package), which result in .fif files (extension *bem-sol.fif). </div>
<div><br></div><div>Using these models we want to continue computing the leadfields and doing source reconstruction in Matlab. For the latter, we have our own customized codes; to get the leadfield, we need some step in between. I'm trying to use fieldtrip functions for this but I get stuck. Here's what I do:</div>
<div><br></div><div style>- Import the bem fif files using the mne toolbox (function: mne_read_bem_surfaces).</div><div style>- Attach the imported boundary data to a vol structure as follows:</div><div style><br></div><div style>
<div><font face="courier new, monospace">[bem] = mne_read_bem_surfaces(bemfilz(subno).name);</font></div><div><font face="courier new, monospace">vol = [];</font></div><div><font face="courier new, monospace">vol.bnd.pnt = bem.rr;</font></div>
<div><font face="courier new, monospace">vol.bnd.tri = double(bem.tris);</font></div><div><font face="courier new, monospace">vol.unit = 'mm';</font></div><div><font face="courier new, monospace">vol.cfg.sourceunits = 'mm';</font></div>
<div><font face="courier new, monospace">vol.type = 'bemcp';</font></div><div><font face="courier new, monospace">vol.cfg.numvertices = <a href="http://bem.np">bem.np</a>;</font></div></div><div><div><br>
</div><div style>- Import the sensor locations using ft_read_sens and convert to mm using ft_convert_units.</div><div style>- Check whether the resulting structures are OK for leadfield computation using ft_prepare_vol_sens. This results in an error "Unsupported volume conductor model for MEG".</div>
<div style><br></div><div style>I also tried ft_read_vol, but for Neuromag this needs the meg-pd toolbox which doesn't run on Windows (the Linux we use for the Neuromag software, in turn, is a virtual machine and doesn't have Matlab).</div>
<div style><br></div><div style>Is there a way around this? Using BEM models for MEG data should in principle be possible, but not using Fieldtrip?</div><div style><br></div><div style>Any suggestions would be much appreciated!</div>
<div style>Thanks in advance,</div><div style><br></div><div style>Joram</div><div><br></div><div><br></div>-- <br>Joram van Driel, MSc.<div>PhD student at the University of Amsterdam</div><div>Department of Psychology, Brain & Cognition</div>
</div></div>