[Eeglablist] questions about statcond and Permutation test

David Groppe dgroppe at cogsci.ucsd.edu
Sun May 22 23:17:59 PDT 2011


On Fri, May 13, 2011 at 5:27 PM, Hector Yamil Vidal Dos Santos
<hvidaldossantos at gmail.com> wrote:
> Hello everyone
> I'm using eeglab to perform a permutation test and I have some questions.
> I'm comparing eeg data acquired under two different conditions
> so
> DATA= {[nelectrodes x timepoints x ntrials under cond 1]  ,  [nelectrodes x
> timepoints x ntrials under cond 2]}
>
> I'm using the statcond command as follow
>
> [t df pvals] = statcond(DATA,'mode','perm','paired','off','naccu', 5000);
>
> and this are my questions:
> What type of permutation test is this command running? Is a Monte Carlo
> permutation test (in which permutations can be use more than ones)? or is it
> a approximate randomization test (in which permutations does not repeat)?
>
> I'm using 5000 permutation, but how can I calculate how many permutations is
> necessary to use depending on the amount of trials?

Hi Hector,
   In Manly's handbook,

Manly, B. F. J. (1997). Randomization, Bootstrap, and Monte Carlo
Methods in Biology
 (2nd ed.). London: Chapman & Hall.

He recommends using 1000 permutations for an alpha level of 5% and
5000 for an alpha level of 1%.  He argues that these numbers should be
generally sufficient to estimate p-values within +/- 2%.   If you want
to be precise about it, Manly describes how you can compute confidence
intervals for the p-values.  If you don't need to be precise about it,
you can run the test and handful of times with a fixed number of
permutations and you can see if the results qualitatively change.

I'm afraid I don't know enough about the mechanics of EEGLAB to answer
the other questions.
    cheers,
        -D

>
> Last question: Do I still need to calculate False Discovery Rate and do
> something about it?
>
> Best,
> Yamil
>
> _______________________________________________
> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to eeglablist-unsubscribe at sccn.ucsd.edu
> For digest mode, send an email with the subject "set digest mime" to
> eeglablist-request at sccn.ucsd.edu
>



-- 
David Groppe, Ph.D.
Postdoctoral Researcher
Kutaslab
Dept. of Cognitive Science
University of California, San Diego
http://www.cogsci.ucsd.edu/~dgroppe/




More information about the eeglablist mailing list