[FieldTrip] Cluster Permutation Data Formatting

[Eric Maris]

Dear Kambiz,


Type "help ft_neighbourselection" on the Matlab command line. This will
provide you with some information. 

Most likely you will decide not to cluster in space (but only in time and
frequency). I this case, the neighbourhood structure of both channels has to
be empty. 


Good luck,

Eric Maris


dr. Eric Maris
Donders Institute for Brain, Cognition and Behavior
Radboud University
P.O. Box 9104
6500 HE Nijmegen
The Netherlands
T:+31 24 3612651
Mobile: 06 39584581
F:+31 24 3616066
mailto:e.maris at donders.ru.nl
http://www.nphyscog.com/



> -----Original Message-----
> From: fieldtrip-bounces at donders.ru.nl [mailto:fieldtrip-
> bounces at donders.ru.nl] On Behalf Of Tavabi, Kambiz
> Sent: maandag 13 december 2010 20:42
> To: nathanweisz at me.com
> Cc: fieldtrip at donders.ru.nl
> Subject: Re: [FieldTrip] Cluster Permutation Data Formatting
> 
> Embarassing, off course I do! I apologize about that. However I am
> still having other downstream errors due to some of the definitions.
> FYI I changed cfg.channel to {['MLT']; ['MRT']}. Running the script
> resulted in
> 
> ??? Undefined function or variable "sens".
> Error in ==> ft_neighbourselection at 95
> if ~isstruct(sens)
> Error in ==> neighbourselection at 17
> [varargout{1:nargout}] = funhandle(varargin{:});
> Error in ==> statistics_wrapper at 241
>     cfg.neighbours = neighbourselection(cfg,varargin{1});
> Error in ==> ft_freqstatistics at 125
> [stat, cfg] = statistics_wrapper(cfg, varargin{:});
> 
> I maybe misunderstanding what is clustered. Is it sensors or adjacent
> time-frequency bins? Because I only have two channels clustering
> adjacent sensors will fail. Perhaps you could advise on whether
> cfg.correctm ='cluster'; is the right choice. Should I consider another
> option? and if so what happens to:
> cfg.clusteralpha     = 0.05;
> cfg.clusterstatistic = 'maxsum';
> cfg.minnbchan        = 2;
> cfg.clustertail      = 0;
> 
> Is there another tutorial which discusses other routines for the
> various multiple comparison correction routines?
> -K.
> -------------------------------------------------------------
> Kambiz Tavabi PhD
> Biomedical Imaging Laboratory
> The Children's Hospital of Philadelphia
> 34th Street and Civic Center Boulevard
> Philadelphia, Pa. 19104
> Tel: 267.426.0302
> email: tavabik at email.chop.edu
> -------------------------------------------------------------
> ________________________________________
> From: fieldtrip-bounces at donders.ru.nl [fieldtrip-bounces at donders.ru.nl]
> On Behalf Of Nathan Weisz [nathanweisz at mac.com]
> Sent: Monday, December 13, 2010 2:12 PM
> To: Email discussion list for the FieldTrip project
> Subject: Re: [FieldTrip] Cluster Permutation Data Formatting
> 
> hi,
> 
> looks like you have a typo in your data structure (lable instead of
> label).
> 
> good luck,
> n
> 
> On 13.12.2010, at 17:33, Tavabi, Kambiz wrote:
> 
> > Greetings, I am using the recommendations on
> http://fieldtrip.fcdonders.nl/tutorial/cluster_permutation_timelock#wit
> hin-subjects_experiments to carry out within-subjects permutation tests
> on MEG data from 10 subjects in a 2x2 design. Briefly, for each subject
> I have the conditional source-space (2 virtual electrodes) time-
> frequency representations computed between 1.5-80Hz (0.5 freq step)
> using Morlet wavelets throughout the stimulus duration (i.e., trial
> epochs from -0.8 - 1.6 sec). As per tutorial I have created data
> structures, which to the best of my knowledge are suitable for use with
> ft_freqstatistics(cfg, cond1, cond2). Also, I have minimally modified
> the routine in the tutorial to create the cfg structure. As the final
> state of the dataset was not arrived at using the FT pipeline I am
> afraid there might be input structure field(s) I may have not defined,
> or defined incorrectly (e.g. channel labling and layout). Essentially I
> simply wish to run the tests and view res
>  ults on two time-frequency representation plots (eg., imagesc)! I
> would very much appreciate if someone with more experience in these
> routines would take a look at my script (see attachment) and e.g. data
> structure and let me know what is missing and/or wrong with them.
> >
> > ExpHi =
> >         freq: [1x158 double]
> >         time: [1x2880 double]
> >    powspctrm: [4-D double]
> >       dimord: 'subj_chan_freq_time'
> >        lable: {2x1 cell} (ie., {['MLT']; ['MRT']}
> >
> > size(ExpHi.powspctrm) =
> >          10           2         158        2880 (i.e., Subjects x
> channes x freqs x timepoints)
> >
> > Currently I am recieving the following errors
> >
> > ??? Reference to non-existent field 'label'.
> > Error in ==> prepare_timefreq_data>forcedimord at 509
> >  output.label  = input.label;
> > Error in ==> prepare_timefreq_data at 87
> >  [remember{c}, hascrsspctrm] = forcedimord(varargin{c}); Error in ==>
> statistics_wrapper at 217
> >  [cfg, data] = prepare_timefreq_data(cfg, varargin{:}); Error in ==>
> ft_timelockstatistics at 117 [stat, cfg] = statistics_wrapper(cfg,
> varargin{:});
> >
> >
> > -K.
> > ------------------------------------------------------------
> > Kambiz Tavabi PhD
> > Biomedical Imaging Laboratory
> > The Children's Hospital of Philadelphia
> > 34th Street and Civic Center Boulevard
> > Philadelphia, Pa. 19104
> > Tel: 267.426.0302
> > email: tavabik at email.chop.edu
> > ------------------------------------------------------------
> >
> >
> <ft_freqstats_script.m>_______________________________________________
> > 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
> 
> _______________________________________________
> fieldtrip mailing list
> fieldtrip at donders.ru.nl
> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip