beamformer on yokogawa data, grad.tra structure

[Sangita Dandekar]

Hi Vladimir and Fieldtrip list,

Thanks for the below reply!   I was wondering if you or anyone familiar with
the yokogawa MEG system
could verify that we are using an appropriate grad.tra matrix and then
subsequently determining the channel leadfield
from grad.tra correctly.   Currently, we are using the generic  definition
for the grad.tra matrix from the yokogawa2grad.m
file in the private fieldtrip directory:

% Define the pair of 1st and 2nd coils for each gradiometer
grad.tra = repmat(diag(ones(1,size(grad.pnt,1)/2),0),1,2);

% Make the matrix sparse to speed up the multiplication in the forward
% computation with the coil-leadfield matrix to get the channel leadfield
grad.tra = sparse(grad.tra);

Each of our channels is an axial gradiometer with two coils so I think that
the above definition should be fine, but
just wanted to check to be sure.

One possibly complicating factor is that MEG160, the software that we use
for data collection with the yokogawa system, has a list of
'calibration weights' for each gradiometer that are determined at each
sensor tuning prior to data collection.  There is one calibration
weight determined per channel (or 1 weight for every pair of coils).  Do
these calibration weights need to be accounted for when
determining grad.tra or the channel leadfield?

Thanks!
Sangi


On Tue, Feb 2, 2010 at 2:07 PM, Vladimir Litvak <v.litvak at ion.ucl.ac.uk>wrote:

> Dear Sangi,
>
> There is no need to convert your data to planar gradient. The
> assumption is that the relation between coils and channels is
> described by the grad.tra matrix. You can look at it and make sure it
> is correct for your system (write back if not). The megplanar function
> as apparent from the error message has explicit support for some
> particular MEG systems and Yokogawa is not one of those. I'm not sure
> how easy it would be to support it generically as there might be
> several variants of Yokogawa systems which can be quite hard to
> distinguish. But for your particular system you can try to implement
> it yourself.
>
> Best,
>
> Vladimir
>
> On Tue, Feb 2, 2010 at 5:22 PM, Sangita Dandekar
> <sangita.dandekar at gmail.com> wrote:
> > Hi,
> > Am hoping to apply beamforming based source localization to MEG data from
> a
> > Yokogawa system.   Think I've managed to coregister MRI and sensor
> > coordinate systems, so that part of the problem is pretty much under
> > control.
> > What I'm wondering about is what the assumptions are of the
> > prepare_leadfield and other source localization scripts about the input
> > gradiometer data.  Haven't looked at it too closely yet, but does it
> assume
> > that the input sensor data is planar gradient data?  If so am assuming
> that
> > inputting the raw data from the Yokogawa system (axial gradiometers) is
> > incorrect?   Or does fieldtrip distinguish between different types of
> > gradiometers using the input .grad structure?
> > I tried to convert the axial gradiometer data from the yokogawa system to
> > planar gradient data by using the megplanar function as shown below, and
> > receive the following error:
> > (Even if it isn't necessary for source localization, it would be nice to
> be
> > able to view the data as planar gradient data)
> >>> cfg=[];
> >>> cfg.planarmethod='sincos';
> >>> megplanar(cfg, righttrials);
> > the input is raw data with 156 channels and 46 trials
> > ??? Error using ==> checkdata at 478
> > This function requires ctf151, ctf275, bti148 or bti248 data as input,
> but
> > you are giving meg data.
> > Error in ==> megplanar at 228
> > data  = checkdata(data, 'datatype', {'raw' 'freq'}, 'feedback', 'yes',
> > 'ismeg', 'yes', 'senstype', {'ctf151', 'ctf275',
> > 'bti148', 'bti248'});
> >>>
> > Some background information:  used the ft yokogawa2grad.m function
> (stored
> > in private FT directory) to create the gradient structure.  Here is what
> > data structure for
> > one set of trials looks like:
> >>> righttrials
> > righttrials =
> >       trial: {1x46 cell}
> >       label: {1x156 cell}
> >        time: {1x46 cell}
> >     fsample: 500
> >        grad: [1x1 struct]
> >      offset: [46x46 double]
> >         cfg: [1x1 struct]
> >>> righttrials.grad
> > ans =
> >       pnt: [314x3 double]
> >       ori: [314x3 double]
> >       tra: [157x314 double]
> >     label: {157x1 cell}
> >      unit: 'cm'
> >>>
> >
> >
> > Thanks in advance for any help!
> > Sangi
> >
> >
> >
> > ----------------------------------
> >
> > The aim of this list is to facilitate the discussion between users of the
> > FieldTrip toolbox, to share experiences and to discuss new ideas for MEG
> and
> > EEG analysis.
> >
> > http://listserv.surfnet.nl/archives/fieldtrip.html
> >
> > http://www.ru.nl/fcdonders/fieldtrip/
>
> ----------------------------------
> The aim of this list is to facilitate the discussion between users of the
> FieldTrip  toolbox, to share experiences and to discuss new ideas for MEG
> and EEG analysis. See also
> http://listserv.surfnet.nl/archives/fieldtrip.html and
> http://www.ru.nl/neuroimaging/fieldtrip.
>

----------------------------------
The aim of this list is to facilitate the discussion between users of the FieldTrip  toolbox, to share experiences and to discuss new ideas for MEG and EEG analysis. See also http://listserv.surfnet.nl/archives/fieldtrip.html and http://www.ru.nl/neuroimaging/fieldtrip.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20100409/bf5408d4/attachment-0001.html>