[Eeglablist] statistics on coherence values

[Marco Congedo]

Dear Stephanics,

 anytime you wish to test a series of univariate hypotheses with EEG data,
the optimal solution is the randomization-permutation approach (known as
t-max test).
The advantages are:

1) It corrects for multiple comparisons keeping the family-wise error rate
below the nominal level (the probability to commit even only one false
positive is less than alpha)

2) It adaptively accounts for the correlation structure of the data,
NO MATTER what that is, hence is more powerful than bonferroni and
relative improvements thereof.

3) It does not assume gaussiantity of data (although, strictly speaking,
if the data is not simmetrically distributed a tranformation to improve
simmetricity is advisable).

This is a good paper about the subject:

 Holmes AP, Blair RC, Watson JD, Ford I.
 Nonparametric analysis of statistic images from functional mapping 
experiments. J Cereb Blood Flow Metab. 1996 Jan;16(1):7-22.

For a primer see

  Nichols TE, Holmes AP.
 Nonparametric permutation tests for functional neuroimaging: a primer
with  examples. Hum Brain Mapp. 2002 Jan;15(1):1-25.

As far as implementations, a free software that does this is in the
LORETA-Key Package
http://www.unizh.ch/keyinst/NewLORETA/LORETA01.htm

I will release a stand-alone software soon...but its not ready.
I code in object Pascal, so I do not have Matlab code
(but I would surely share my low-level pascal code).

Livio Finos, at University of Ferrara (Italy) may be interested
in implementing the test in Matlab for EEGlab.
He has already several routines written in matlab.
His e-mail is lfinos at stat.unipd.it

Hope this help

Marco


-- 
Marco Congedo, PhD
SIAMES Project, IRISA
Campus de BeauLieu
35042 Rennes France
Home Page:
http://www.irisa.fr/siames/GENS/mcongedo/MC_Home.html




> Dear All,
>
> I am analyzing channel cross coherence for 32 channel eeg data, I have the
> resulting matrices of the crossf function for all the possible channel
> combinations with a 0.05 bootsrap significance mask applied on the
> prestimulus
> baseline. My question is what is the correct statistical procedure to
> compare
> different experimental conditions? In many papers I have read that
> coherence
> values were Fisher-z-transformed before paired Wilcoxon tests on
> individual
> channel-pairs or before ANOVA for a number of channel pairs. So my above
> question is manifold:
> - Is it necessary to apply Fisher-z-transformation on my data?
> - Which statistical test is the optimal? Because if I do the Wilcoxon
> test,
> the significance levels are not corrected for multiple comparisons, but if
> I
> make ANOVA I can test only a subset of the channel pairs to avoid loss of
> statistical power of the ANOVA. Is there any way to avoid this trap?
>
> Thanks for any help and advice in advance!
>
>
> Gabor Stefanics
>
> _______________________________________________
> eeglablist mailing list eeglablist at sccn.ucsd.edu
> http://sccn.ucsd.edu/mailman/listinfo/eeglablist
> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to
> eeglablist-unsubscribe at sccn.ucsd.edu
>


-- 
Marco Congedo, PhD
SIAMES Project, IRISA
Campus de BeauLieu
35042 Rennes France
Home Page:
http://www.irisa.fr/siames/GENS/mcongedo/MC_Home.html