Freqstatistics Yields Zero Significant Clusters?

[Saskia Haegens]

Hi Charles,

The reason that clusterplot gives an error is because you do not average
over frequencies (which is required for clusterplot). So you should either
set cfg.avgoverfreq='yes' when running freqstatistics, or plot the results
in another way (e.g. use singleplot or feed only one freqbin at a time into
clusterplot).
Hope this helps!

Best,
Saskia

> -----Original Message-----
> From: FieldTrip discussion list [mailto:FIELDTRIP at NIC.SURFNET.NL] On
> Behalf Of Charles Cook
> Sent: 08 June 2009 21:52
> To: FIELDTRIP at NIC.SURFNET.NL
> Subject: Re: [FIELDTRIP] Freqstatistics Yields Zero Significant Clusters?
>
> Hi Eric,
>
> Yes that did seem to a further unusually high number, especially since
> they're all standard-81! Let me give you an idea of where our code is
> presently at:
>
> ============================================
>
> cfg = [];
> cfg.neighbourdist    = 4;
> cfg.elec             = elec;
> cfg.statistic        = 'indepsamplesT';
> cfg.minnbchan        = 0;
> cfg.clusteralpha     = 0.05;
> cfg.alpha            = 0.05;
> cfg.clustertail      = 0;
> cfg.numrandomization = 5000;
>
> cfg.latency          = [250 500];
> cfg.frequency        = [4 7];
> cfg.avgovertime      = 'no';
> cfg.avgoverfreq      = 'no';
> cfg.avgoverchan      = 'no';
>
> cfg.correctm         = 'cluster';
> cfg.method           = 'montecarlo';
> cfg.feedback         = 'gui';
> cfg.design           = [1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
> 21 22;   % subject number
>                         1 1 1 1 1 1 1 1 1 1  1  2  2  2  2  2  2  2  2  2
> 2
>  2]; % condition number
>
> cfg.uvar = 1;                                   % "subject" is unit of
> observation
> cfg.ivar = 2;                                   % "condition" is the
> independent variable
> stat = freqstatistics(cfg, malefeat_all, femloc_all);
>
> %cfg = [];
> %[freq_maleloc] = freqdescriptives(cfg, maleloc_all);
> %[freq_femloc]  = freqdescriptives(cfg, femloc_all);
>
> cfg = [];
> cfg.zlim = [-6 6];
> cfg.alpha = 0.05;
> clusterplot (cfg, stat); <----still failing here
>
> ============================================
>
> Reading power on 81 channels
> not computing grand average, but keeping individual power for 11 subjects
> not computing grand average, but keeping individual power for 11 subjects
> selected 81 channels
> selected 6 time bins
> selected 4 frequency bins
> Warning: PACK can only be used from the MATLAB command line.
> > In fieldtrip\private\prepare_timefreq_data at 310
>   In fieldtrip\private\statistics_wrapper at 206
>   In freqstatistics at 132
>   In CMCWM2_std81_junk_with at 144
> Obtaining the electrode configuration from the configuration.
> there are on average 83.0 neighbours per channel
> using "statistics_montecarlo" for the statistical testing
> using "statfun_indepsamplesT" for the single-sample statistics
> constructing randomized design
> total number of measurements     = 22
> total number of variables        = 2
> number of independent variables  = 1
> number of unit variables         = 1
> number of within-cell variables = 0
> number of control variables      = 0
> using a permutation resampling approach
> repeated measurement in variable 1 over 22 levels
> number of repeated measurements in each level is 1 1 1 1 1 1 1 1 1 1 1 1 1
> 1
> 1 1 1 1 1 1 1 1
> computing a parmetric threshold for clustering
> estimated time per randomization is 0 seconds
> found 2 positive clusters in observed data
> found 1 negative clusters in observed data
> using a cluster-based method for multiple comparison correction
> the returned probabilities and the thresholded mask are corrected for
> multiple comparisons
> ??? Assignment has more non-singleton rhs dimensions than non-singleton
> subscripts
>
> Error in ==> clusterplot at 91
>     sigposCLM(:,:,iPos) = (posCLM == sigpos(iPos));
>
> Error in ==> CMCWM2_std81_junk_with at 153
> clusterplot (cfg, stat);
>
> ============================================
>
> Should our cfg.neighbourdist have a higher number if we are using a
> Standard-81 layout? We've tried a few different variations with that
> number
> and still ended up with the same error.
>
> Cheers,
>
> Charles
>
> On Mon, 8 Jun 2009 17:00:22 +0200, Eric Maris <e.maris at DONDERS.RU.NL>
> wrote:
>
> >Hi Charles,
> >
> >83.0 neighbours per channel does not make sense. For EEG-channels this
> >number typically is 4 and for MEG-channels it is typically 6. Have a look
> at
> >the neighbourhood geometry structure that is constructed by
> freqstatistics.
> >I guess this structure is far too wide (a channel is considered a
> neighbour
> >of almost every other channel).
> >
> >
> >Best,
> >
> >Eric
> >
>
> ----------------------------------
> 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/neuroimaging/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/neuroimaging/fieldtrip.