Saving preprocessed data in a raw MEG format

[Vladimir Litvak]

Dear all,

Since this thread is turning to be about promoting other software
packages, I'd just like to mention that SPM8 has imaging source
reconstruction for evoked fields that depending on the settings can
behave similarly to minimum norm or to LORETA or to a more focal
methods called MSP. Since SPM8 and Fieldtrip are very tightly
integrated (see http://fieldtrip.fcdonders.nl/integrating_with_spm8),
your Fieldtrip data can be converted to SPM8 format quite easily. In
SPM8 you can also do statistics on resulting images.

Best,

Vladimir

On Wed, Jul 14, 2010 at 9:39 AM, Geoff Brookshire
<Geoff.Brookshire at mpi.nl> wrote:
> Hi,
>
> Thanks to everyone for a lot of helpful advice!  I'm going to go through and
> try a lot of this, and I'll report back to the list when I have something
> that works.
>
> I'll briefly address a couple of suggestions.
> Unfortunately, we've done some pretty heavy-duty preprocessing: rejecting
> lots of artifacts, truncating trials, looking at stimulus-response
> combinations, and correcting some errors in the parallel port output
> channel.  I'm afraid I'll be chained to the complicated solutions, then.
>
> Second, and in regards to using the source localization algorithms that
> Freesurfer already has -- I'm a novice in this area, so I'm taking the
> conclusions of O. Jensen and C. Hesse [Estimating distributed
> representations of evoked responses and oscillatory brain activity. In MEG
> An Introduction to Methods" (edited by Salmelin, Kringleback and Hansen).
> Oxford University Press].  They advise against using beamforming for ERFs:
>>
>> A further practical requirement [for beamforming] is that the estimate of
>> the data cross-covariance matrix is accurate and that
>> the cross-covariance matrix is invertible. These are the problems often
>> encountered when applying
>> beamformingfor source analysis of ERFs. The cross-covariance matrix can
>> become rank-deficient (and
>> hence invertible) in two circumstances: if the epoch of interest contains
>> fewer time samples than sensors, or
>> (at least theoretically) if the time-locked averaged MEG signal reflects
>> the activity of only a very small
>> number of neuronal sources, and any background activity and noise
>> effectively disappears by averaging an
>> extremely large number of trials. In practice, rank-deficient
>> cross-covariance matrices can be made
>> invertible by regularization, which effectively involves injecting
>> additional noise into the signal, and this, in
>> turn, tends to spatially smear the neuronal activity. For this reason,
>> beamforming is not widely used in
>> MEG/EEG source analysis of evoked activity.
>
> Updates to follow!
>
> -geoff
>
>
> Nathan Weisz wrote:
>>
>> hi,
>>
>>> I'm using MNE software to perform the localization, since Fieldtrip is
>>> only able to do this on time-frequency data now.  The thing is that I've
>>
>> I do not think that this is correct. you may want to look at Figure 2E in:
>> Grützner et al. (2010). Neuromagnetic correlates of perceptual closure
>> processes. J Neurosci
>>
>> at least the authors state that they used fieldtrip, so i assume they used
>> minumumnormestimate.m.
>>
>> it might be correct that this function is not so intensively developed, as
>> not so many fieldtrippers seem to use it.
>>
>> but before going through all the pain of exporting your data, you may want
>> to quickly check if your results make sense using the FT-version (e.g.
>> sanity checks on some ERP where you know where the main generators should
>> be).
>>
>> in case you are not satisfied with the FT-minimumnorm, an alternative to
>> MNE would be to use brainstorm. i think they use mat-files, so it should be
>> fairly easy to mimick to output of some brainstorm processing stage and
>> continue with source analysis. haven't tried it myself but i have a
>> colleague who does this quite successfully.
>>
>> yet another alternative would be to use nutmeg. it should contain a
>> sLORETA option and a function to directly read FT-preproc-structures. at
>> least the developer wrote such a function to read some of my FT-structures
>> into nutmeg. so if it is not contained in the release you might ask to
>> become a beta-tester :-)
>>
>> good luck,
>> n
>>
>>
>>
>> *--------------------------------------------*
>> Dr. Nathan Weisz
>>
>> OBOB-Lab
>> University of Konstanz
>> Department of Psychology
>> P.O. Box D23
>> 78457 Konstanz
>> Germany
>>
>> Tel: ++49 - (0)7531 - 88 45 84
>> Email: nathan.weisz at uni-konstanz.de <mailto:nathan.weisz at uni-konstanz.de>
>> Homepage: http://www.uni-konstanz.de/obob
>>
>> "Nothing shocks me. I'm a scientist." (Indiana Jones)
>>
>> ----------------------------------
>>
>> 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.