[FieldTrip] Using fieldtrip statistic for cluster analysis of an already existing dataset

[Lustenberger Caroline]

Dear Steve
 
Thank you very much, I was able to calculate montecarlo statistics with
the fieldtrip toolbox using eeglab data. 
Probably you can help me answering this question too:
 
How can I define neighbour-timepoints, normally all tutorials explain
how to define neighbour sensors or channels, but have only 1 channel and
185 timepoints. Thus, I want to perform cluster analysis to see whether
a certain timepoint cluster is significant.
 
Thanks again and best wishes
Caroline
 
 

________________________________

Von: fieldtrip-bounces at donders.ru.nl
[mailto:fieldtrip-bounces at donders.ru.nl] Im Auftrag von Stephen
Politzer-Ahles
Gesendet: Mittwoch, 28. September 2011 16:37
An: fieldtrip at donders.ru.nl
Betreff: Re: [FieldTrip] Using fieldtrip statistic for cluster analysis
of an already existing dataset


Caroline,

This should be possible (I recently did something very similar, although
I was analysizing ERPs rather than ERSPs).

First you need to export the data for each subject from EEGLAB to
Fieldtrip using EEGLAB's function eeg2fieldtrip(); when I did this I
specified the option 'timelockanalysis' to compute ERPs of the EEGLAB
data, but in  your case I guess you will want to use 'freqanalysis'
instead.

Then, you need to compute grand averages using ft_freqgrandaverage()
with cfg.keepindividual='yes'. Then you will be ready to perform cluster
statistics following the instructions at
http://fieldtrip.fcdonders.nl/tutorial/cluster_permutation_freq .

Best,
Steve Politzer-Ahles




	Message: 1
	Date: Wed, 28 Sep 2011 10:46:16 +0200
	From: "Lustenberger Caroline"
<Caroline.Lustenberger at kispi.uzh.ch>
	To: <fieldtrip at donders.ru.nl>
	Subject: [FieldTrip] Using fieldtrip statistic for cluster
analysis of
	       an      already existing dataset
	Message-ID:
<8D58B6880DD231469FA0128331363CED03DB01AB at exzh1.kispi.int>
	Content-Type: text/plain; charset="us-ascii"
	
	Dear all
	
	Using EEGlab I have calculated ITC and ERSP of 16 subjects of 1
channel
	and the frequency range 1-4Hz (in average), thus I have 2 data
arrays:
	
	1st Condition Sham: 16 x 185 --> 16=number of subjects,
185=timepoints
	with event-related spectral values or intertrial coherence
values
	2nd Condition Field: 16 x 185
	
	To see if there is any difference between the 2 conditions in
all
	timepoints I performed 185 paired t-tests and now I should
control for
	multiple comparison using a cluster analysis (meaning find out
the
	probability that a certain number of neighbour-timepoints are
	suprathreshold).
	
	Is there a possibility to do that using fieldtrip statistics?
And how
	can I do that?
	
	Thank you very much for your help
	Caroline
	-------------- next part --------------
	An HTML attachment was scrubbed...
	URL:
<http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20110928/a
27bc4a9/attachment-0001.html>
	
	------------------------------
	
	_______________________________________________
	fieldtrip mailing list
	fieldtrip at donders.ru.nl
	http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
	
	End of fieldtrip Digest, Vol 10, Issue 44
	*****************************************
	





-- 
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/20110928/8eb069df/attachment.html>