[FieldTrip] Beamformer: different length of baseline and post baseline interval

"Jörn M. Horschig" jm.horschig at donders.ru.nl
Mon Oct 29 10:27:14 CET 2012


Hi Akiko,

of course you can use it, but freqanalysis will pad zeros thereby 
increasing the spectral resolution but not add any data specifically. In 
the end, your result will be a more smoothed spectral estimate for the 
padded trials, which allows you to average across trials with different 
length (since they the spectral resolution is now equal across trials) 
or in this case contrast conditions of different lengths. I don't see 
much sense in doing that for the baseline, apart from tricking the 
algorithm, but from a more pragmatic point of view: if no reviewer 
complaints then it's fine ;) I bet there are some other consequences 
that I don't foresee; if someone knows, please let me/us know :)

Best,
Jörn

On 10/26/2012 4:52 PM, Akiko Ikkai wrote:
> Hi,
>
> I know this is not a recent post, but please allow me to ask a 
> follow-up question; can you use cfg.pad for ft_freqanalysis for the 
> baseline in this case (assuming post-baseline period of interest is 
> uniform length, say 1000ms)?
>
> Such as:
> % extract baseline (500ms)
> cfg             = [];
> cfg.toilim      = [-.7 -.2];
> data_BL         = ft_redefinetrial(cfg,ft_data);
>
> % pad to make it 1000ms and run ft_freqanalysis
> cfg             = [];
> *cfg.pad         = 1;*
> cfg.method      = 'mtmfft';
> cfg.output      = 'powandcsd';
> cfg.foi         = foi;
> cfg.taper       = 'dpss';
> cfg.tapsmofrq   = smooth;
> freq_BL         = ft_freqanalysis(cfg,data_BL);
>
> Thanks! Akiko
>
> On Mon, Jul 2, 2012 at 4:07 PM, Johanna Zumer 
> <johanna.zumer at donders.ru.nl <mailto:johanna.zumer at donders.ru.nl>> wrote:
>
>     Dear Anna,
>
>     Ideally for the common filter, you want the same amount of data
>     T(s) per condition, where T = N x tw (and N is number of trials
>     and tw is timewindow length).  In your case, if the baselines for
>     each conditions can be combined into one general baseline, and if
>     you happen to have 100 trials per condition, then T_baseline = 3 x
>     100 x 0.5s = 150s.  If you then use 1.5s length post-baseline,
>     then T_each_condition = 100 x 1.5s = 150s, so you now have equal T
>     for each condition for the common filter.
>
>     However, in order to have an equal effect of tapers and
>     edge-effects on the different conditions, you should use equal
>     time window lengths in freqanalysis.  Thus it would be better to
>     split your post-baseline data into 3 segments of 500ms each before
>     calling ft_freqanalysis, which again gives T = 100 x 0.5 x 3 = 150s.
>
>     Cheers,
>     Johanna
>
>     2012/6/29 Anna Wilsch <wilsch at cbs.mpg.de <mailto:wilsch at cbs.mpg.de>>
>
>
>         Dear Fieldtrippers,
>
>         I'm trying to beamform my MEG data by building a common filter
>         including three conditions and a baseline for each condition.
>         The baseline intervals have a duration of 500 ms. I was
>         wondering if it is ok if the post-baseline data are longer
>         than that (1000 - 2000 ms). Does it have any negative impact
>         on the cross-spectral-density matrix and/or the common filter?
>         Would that still be a valid operation to do or is it necessary
>         that baseline and post baseline data have the same length?
>         Thank you for your comments.
>
>         Cheers,
>         Anna
>
>
>         Anna Wilsch, Dipl.-Psych.
>         Auditory Cognition Research Group
>         Max Planck Institute for Human Cognitive and Brain Sciences
>         Stephanstr. 1a - Leipzig, Germany
>         (p) +49 (0)341 9940 2641 <tel:%2B49%20%280%29341%209940%202641>
>         (e) wilsch at cbs.mpg.de <mailto:wilsch at cbs.mpg.de>
>
>         _______________________________________________
>         fieldtrip mailing list
>         fieldtrip at donders.ru.nl <mailto:fieldtrip at donders.ru.nl>
>         http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>
>
>
>     _______________________________________________
>     fieldtrip mailing list
>     fieldtrip at donders.ru.nl <mailto:fieldtrip at donders.ru.nl>
>     http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>
>
>
>
> -- 
> Akiko Ikkai, Ph.D.
> Postdoctoral Fellow
> Department of Psychological and Brain Sciences
> Johns Hopkins University
> Ames Hall, 3400 N. Charles St.
> Baltimore, MD 21218
>
>
>
>
> _______________________________________________
> fieldtrip mailing list
> fieldtrip at donders.ru.nl
> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip


-- 
Jörn M. Horschig
PhD Student
Donders Institute for Brain, Cognition and Behaviour
Centre for Cognitive Neuroimaging
Radboud University Nijmegen
Neuronal Oscillations Group
FieldTrip Development Team

P.O. Box 9101
NL-6500 HB Nijmegen
The Netherlands

Contact:
E-Mail: jm.horschig at donders.ru.nl
Tel:    +31-(0)24-36-68493
Web: http://www.ru.nl/donders

Visiting address:
Trigon, room 2.30
Kapittelweg 29
NL-6525 EN Nijmegen
The Netherlands

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20121029/b8889e9e/attachment.html>


More information about the fieldtrip mailing list