[FieldTrip] Covariance matrix, timelocking and lcmv beamformer

[Kaelasha Tyler]

Hi Eelke,

Thanks so much for this.

As it happens I copied the script recommending to time lock twice some time ago, and wasn't able to see it when I looked through the FT site this time. 
So, problem solved all round.

Cheers!
Kaelasha
________________________________________
From: fieldtrip-bounces at science.ru.nl [fieldtrip-bounces at science.ru.nl] on behalf of Eelke Spaak [eelke.spaak at donders.ru.nl]
Sent: Thursday, 28 August 2014 12:12 AM
To: FieldTrip discussion list
Subject: Re: [FieldTrip] Covariance matrix, timelocking and lcmv beamformer

Dear Kaelasha,

The tutorial is correct, it's best to use the averaged covariance,
computed on the unaveraged data (so a single call to
ft_timelockanalysis). We thought we had removed the example script you
mentioned some time ago, but apparently somewhere on the wiki it still
lingers. Could you let us know where you found this? Or better yet,
just remove it? (It is a wiki, after all.)

Best,
Eelke

On 26 August 2014 02:57, Kaelasha Tyler <ktyler at swin.edu.au> wrote:
> Hi all,
>
> I am raking through my script trying to sort out potential faults in my
> analysis.
> One issue I have come across is whether ft_timelock analysis should be
> performed once, or twice when using lcmv method of beamformer.
>
> The tutorial on lcmv beamfomer found here
> http://fieldtrip.fcdonders.nl/tutorial/beamformer_lcmv?s[]=covariancewindow
> seems to indicate that time-locking is only performed once only in the
> analysis pipeline.
>
> However, elsewhere on the FT site, I found the following advice:
> "In order for the beamformer to work, we need the covariance between sensor
> pairs. This should be the covariance of (a particular latency window) of the
> averaged single trials. This is essentially different from the average
> covariance of the single trials. Therefore, we have to call timelockanalysis
> more than once: the first time we obtain the evoked field, and the second
> time we obtain the covariance matrix/ces at interesting latency/ies"
>
> Which is correct?
>
> I have been time locking twice: using data sets which were already
> time-locked using ft_timelockanalysis, and then time locking again, prior to
> using ft_sourceanalysis........... Wrong?
>
> The phrase 'sanity check' comes up quite often on this list, and I
> absolutely understand! As always, any help is very much appreciated!!
>
> Kaelasha
>
>
>
>
>
>
>
> _______________________________________________
> fieldtrip mailing list
> fieldtrip at donders.ru.nl
> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
_______________________________________________
fieldtrip mailing list
fieldtrip at donders.ru.nl
http://mailman.science.ru.nl/mailman/listinfo/fieldtrip