clusterrandanalysis actvsblT
Eric Maris
maris at NICI.RU.NL
Tue Jul 31 17:57:53 CEST 2007
Dear Andre,
> Yes, I understand about the pvalues not working for comparing within a
trial
> (from baseline to activation). But what about comparing across different
> periods of time, using independent samples? Is it okay to set avgovertime
=
> 'yes' then?
As long as you average over periods of time of equal length (in different
experimental conditions), it is okay.
>
> For example, I am comparing two conditions in a ssvep paradigm: attend
left
> and attend right, for one subject. I am not quite sure what I am doing
> wrong, as I am sure that there are differences between these two
conditions,
> although the clusters that I get are all insignificant.
>
> In this case, when the subject attends left they attend to 7.5 Hz, and
when
> they attend right they attend to 10 Hz flicker stimulus. I am expecting
that
> these frequencies, as well as higher harmonics will be sensitive to
> attention (as previously reported in the ssvep lit). However, my clusters
> all have p values above 0.2. I tried this with avgovertime set to both yes
> and no, and still no sig clusters.
>
> Why are the insignificant clusters even returned at all, especially when
> there is a parameter value for determining significant clusters,
> alphathresh, which is set in the cfg but never used? I know that other
> people have had the same issue before...
It is a misunderstanding that cfg.alphathresh is never used. It is used to
define the clusters. If you change the value of cfg.alphathresh you get
different clusters. Be aware that cfg.alphathresh must be chosen independent
of the data if you want to control the FA rate.
Kind regards,
Eric Maris
>
> Thanks again,
>
> Andre
>
> Here is my config (I am using just the 29 parietal and occipital
electrodes
> for this analysis).
>
> elec.pnt = elec.pnt(1:29,:);
> elec.label = elec.label(1:29,:);
>
> cfg=[];
> cfg.statistic = 'indepsamplesT';
> cfg.alphathresh = 0.05;
> cfg.makeclusters = 'yes';
> cfg.minnbchan = 2;
> cfg.clusterteststat = 'maxsum';
> cfg.onetwo = 'twosided';
> cfg.alpha = 0.05;
> cfg.nranddraws = 500;
> cfg.channel = 'all'; %TFR(1).elec.label(1:64); %{'EEG'};
> cfg.latency = [1 5 ];
> cfg.elec = elec;
> cfg.neighbourdist = 49;
> cfg.avgovertime = 'no';
>
> [clusrand] = clusterrandanalysis(cfg, TFR1, TFR2);
>
>
> -----Original Message-----
> From: FieldTrip discussion list [mailto:FIELDTRIP at NIC.SURFNET.NL] On
Behalf
> Of Eric Maris
> Sent: Monday, July 30, 2007 3:42 PM
> To: FIELDTRIP at NIC.SURFNET.NL
> Subject: Re: [FIELDTRIP] clusterrandanalysis actvsblT
>
> Dear Andre,
>
> > On a related note, I would like to ask about the clusterrandanalysis
> option
> > to set avgovertime = 'yes'
> >
> > In this case, as far as I am aware, the frequency-space-time-trial
> quadlets
> > which clusterrand analysis uses become doublets of frequency-space. In
> this
> > case, wouldn't it be possible to compare data periods of different
> lengths,
> > as clusterrandanalysis would just average across your baseline period as
> set
> > 1 and average across the activation period as set 2?
> >
> > Does the distribution of activation over time get factored in to the
> > clusterrandanalysis function even when avgovertime is set to 'yes'?
>
>
> If you try to do this, you should get an error message. Have a try, and if
> you don't get an error message let me know. Anyhow, even without an error
> message, the p-value that you get out of clusterrandanalysis only makes
> sense if the baseline and activation periods are of equal length.
>
>
> Good luck,
>
> Eric Maris
>
>
>
> dr. Eric Maris
>
> NICI/Biological Psychology and
>
> F.C. Donders Center for Cognitive NeuroImaging
>
> University of Nijmegen
>
> P.O. Box 9104
>
> 6500 HE Nijmegen
>
> The Netherlands
>
> T:+31 24 3612651 (NICI)
>
> T:+31 24 3610754 (FCDC)
>
> F:+31 24 3616066 (NICI)
>
> E: maris at nici.ru.nl
>
> MSc Cognitive Neuroscience :www.ru.nl/master/cns/
>
>
>
>
>
>
>
> >
> > Thanks,
> >
> > -Andre
> >
> >
> > -----Original Message-----
> > From: FieldTrip discussion list [mailto:FIELDTRIP at NIC.SURFNET.NL] On
> Behalf
> > Of Eric Maris
> > Sent: Monday, July 30, 2007 1:50 PM
> > To: FIELDTRIP at NIC.SURFNET.NL
> > Subject: Re: [FIELDTRIP] clusterrandanalysis actvsblT
> >
> > Dear Claudia,
> >
> >
> >
> > > i have a questions about clusteranalysis comparing baseline and
> > > activation period. is it really impossible to compare two periods of
> > > different length? and why? as far as i understand the mean of the
> > > baseline is used and so you have only one value per frequency anyway.
> > > can i avoid this problem or do i have to divide my activation period
> > > into different segments because my baseline period is much shorter?
> >
> >
> > You are one of many to complain about this feature of
clusterrandanalysis.
> > Nevertheless, you do have to divide your activation period into
different
> > segments. The reason for this has to do with the null hypothesis that is
> > tested when you use the activation-versus-baseline statistic. This null
> > hypothesis is the following: The data in the baseline period are drawn
> from
> > the same probability distribution (over the spatiotemporal domain) as
the
> > data in the activation period. This null hypothesis only makes sense if
> the
> > spatiotemporal dimensionality of the activation and the baseline period
> are
> > equal.
> >
> >
> > Kind regards,
> >
> > Eric Maris
> >
> >
> >
> >
> >
> >
> >
> >
> > dr. Eric Maris
> >
> > NICI/Biological Psychology and
> >
> > F.C. Donders Center for Cognitive NeuroImaging
> >
> > University of Nijmegen
> >
> > P.O. Box 9104
> >
> > 6500 HE Nijmegen
> >
> > The Netherlands
> >
> > T:+31 24 3612651 (NICI)
> >
> > T:+31 24 3610754 (FCDC)
> >
> > F:+31 24 3616066 (NICI)
> >
> > E: maris at nici.ru.nl
> >
> > MSc Cognitive Neuroscience :www.ru.nl/master/cns/
> >
> >
> >
> >
> > > thanks in advance
> > > claudia
> > >
> > >
> > > --
> > > Pflichtangaben gemäß Gesetz über elektronische Handelsregister und
> > > Genossenschaftsregister sowie das Unternehmensregister (EHUG):
> > >
> > > Universitätsklinikum Hamburg-Eppendorf
> > > Körperschaft des öffentlichen Rechts
> > > Gerichtsstand: Hamburg
> > >
> > > Vorstandsmitglieder:
> > > Prof. Dr. Jörg F. Debatin (Vorsitzender)
> > > Dr. Alexander Kirstein
> > > Ricarda Klein
> > > Prof. Dr. Dr. Uwe Koch-Gromus
> >
> > ----------------------------------
> > 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/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/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/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/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/fcdonders/fieldtrip.
More information about the fieldtrip
mailing list