[FieldTrip] question on beamformers

[Stephen Whitmarsh]

Dear Marcel,

My two cents are the following:

I would recommend to compute common filter over ALL conditions, pre AND
post. This will make sure you use as much information as possible to
calculate your cross-spectral density matrix and subsequent filter. I assume
you assume that the location/orientation of the sources do not change before
and after stimulus. Also, in the case that you are looking for a
post-stimulus beta DECREASE it will help you to accurately determine
post-stimulus activity based on a filter calculated from of a period where
there is most (beta) signal.

Then, for your contrast project your conditions, pre and post *separately*
and do your relative contrast then: e.g. projected post-stimulus activity A
/ projected pre-stimulus activity A, and same for B. Then compare relative A
- or / by B.

Bottom line, do your common filter on everything, project and compare
everything separately after that.

Hope that helps,



Stephen

On 24 January 2011 11:44, Bastiaansen, Marcel <Marcel.Bastiaansen at mpi.nl>wrote:

> Dear all,
>
> I have the following conceptual issue with using beamforming:
>
> I have data from 2 conditions, A and B. Both contain a prestimulus interval
> and a poststimulus interval.
>
> When plotting the TF representations of power, using a *relative* baseline
> (poststimulus interval / prestimulus interval), a strong difference emerges
> between the conditions in the beta frequency range. Especially if I plot the
> *difference between the conditions*, the beta effect is long-lasting,
> narrow-band and topographically focused. Ideal for beamforming, I'd say.
>
> In setting up the beamformer analysis, I am faced with the following issue:
> the effect is most clearly present in the *relative* power changes, so I
> would want to compute spatial filters for both the pre and the post time
> intervals. My hunch would be to compute common spatial filters for both pre
> intervals (preA and preB), and common filters for both post intervals (postA
> and postB). Then, (after projecting the trials through the filters), compute
> the relative power again (pre/post for each condition). Does this approach
> make sense?
>
> Next question is then, at which point in the analysis pipeline should one
> do the condition subtraction? (I'd want to do that because at the TFR level
> the effect comes out most clearly in the condition difference)
>
>
> Best,
> Marcel
>
>
>
> --
> dr. Marcel C.M. Bastiaansen.
>
> Max Planck Institute for Psycholinguistics
> Visiting Adress: Wundtlaan 1, 6525 XD Nijmegen, the Netherlands
> Mailing adress: P.O. Box 310, 6500 AH Nijmegen, the Netherlands
> phone: +31 24 3521 347
> fax:     +31 24 3521 213
> mail: marcel.bastiaansen at mpi.nl
> http://www.mpi.nl/people/bastiaansen-marcel
>
> and
>
> Donders Institute for Brain, Cognition and Behaviour
> Centre for Cognitive Neuroimaging
> Visiting address: Kapittelweg 29, 6525 EN Nijmegen, the Netherlands
> Mailing address: PO Box 9101, 6500 HB Nijmegen, the Netherlands
> phone: + 31 24 3610 882
> fax:   + 31 24 3610 989
> mail: marcel.bastiaansen at fcdonders.ru.nl
> web:
> http://www.ru.nl/neuroimaging/staff/neurocognition_of/marcel_bastiaansen/
> --
> _______________________________________________
> fieldtrip mailing list
> fieldtrip at donders.ru.nl
> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20110124/39673d0d/attachment-0002.html>