[FieldTrip] Question of implementing Fieldtrip

[Stephen Politzer-Ahles]

Hi Jinghao,

You can use cfg.latency to limit the time period in which the test will
look for clusters (for example, if you expect a difference to show up in a
particular small time window and you are only interested in that
difference, and don't want to test earlier or later differences); the
tutorial has some more information. The choice of time window should be
made based on predictions you had before the experiment, not based on
inspection of the data. But like Roemer points out, it's not possible to
save "only clusters with psychological meaning", as the statistical test
doesn't know whether or not clusters have psychological meaning.

Best,
Steve


> Message: 2
> Date: Wed, 19 Sep 2012 11:20:12 +0200
> From: Roemer van der Meij <r.vandermeij at donders.ru.nl>
> To: Email discussion list for the FieldTrip project
>         <fieldtrip at donders.ru.nl>
> Cc: Jinghao Lu <jhlu at ucdavis.edu>
> Subject: Re: [FieldTrip] Question of implementing Fieldtrip
> Message-ID:
>         <
> CA+WpQ37dfrEoQV4gLuVTHWbMrBKexo8HFRprSJ9d5t8w45ZvYg at mail.gmail.com>
> Content-Type: text/plain; charset="iso-8859-1"
>
> Hi Jinghao,
>
> I forwarded your question to the Fieldtrip mailing list. I don't entirely
> understand what you mean only saving clusters with psychological meaning.
> Given the parameters that you chose, you get a certain number of clusters,
> all of which are significant as a whole. In analogy, this means that the
> whole 'mountain range' is significant or not, not the 'individual peaks'.
> You could of course focus on a few of those clusters in your report, as
> long as you mention the other clusters. You could also try to set your
> parameters a little more stringent (i.e. a lower cfg.clusteralpha), this
> would 'reduce the number of peaks in your mountain range', as some peaks
> would then fall below the threshold. Hopefully, your clusters with
> 'psychological meaning' would still show up, albeit a little
> smaller/shorter.
>
> Hope this helps,
> Roemer
>
>
>
>
> > Well in fact I'm doing an intracranial EEG project currently where the
> data
> > should be analyzed statistically. The permutation analysis seems a good
> > choice and first I ran it with sample-specific independent t-test to get
> a
> > rough idea of the data in different conditions.
> >
> > I'm wondering how to apply cluster-based permutation in single sensor
> > analysis where I would like to save only the time clusters with
> > psychological meaning. I could not find the guidance of how to do this in
> > Fieldtrip. Could you please share with me the approach? That'll be really
> > helpful.
> >
> > Best Regards,
> > Jinghao Lu
> >
> >
> > Dept. of Physics
> > Zhejiang University
> > 866 Yuhangtang Road
> > Hangzhou
> > 310058 ZHEJIANG
> > CHINA
> >
> > Temporary Affiliate
> > Dynamic Memory Lab
> > UC Davis
> >
> >
> > Jinghao Lu
> >
> >
> > Dept. of Physics
> > Zhejiang University
> > 866 Yuhangtang Road
> > Hangzhou
> > 310058 ZHEJIANG
> > CHINA
> >
> > Temporary Affiliate
> > Dynamic Memory Lab
> > UC Davis
> >
>
>
>
> --
> Roemer van der Meij M.Sc.
> PhD student
> Donders Institute for Brain, Cognition and Behaviour
> Centre for Cognition
> P.O. Box 9104
> 6500 HE Nijmegen
> The Netherlands
> Tel: +31(0)24 3655932
> E-mail: r.vandermeij at donders.ru.nl
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <
> http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20120919/bc3080b9/attachment-0001.html
> >
>
> ------------------------------
>
> _______________________________________________
> fieldtrip mailing list
> fieldtrip at donders.ru.nl
> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>
> End of fieldtrip Digest, Vol 22, Issue 35
> *****************************************
>



-- 
Stephen Politzer-Ahles
University of Kansas
Linguistics Department
http://www.linguistics.ku.edu/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20120919/0d9cf972/attachment.html>