# [Eeglablist] Get statistical parameters of power spectrum analysis

John, Alexander alexjohn at uni-mainz.de
Thu Nov 30 02:10:07 PST 2017

```Hello everyone,

I'd like to receive exact numeric p-values as well as the other statistic parameters of the power spectrum analysis, for every single electrode or at least for the electrodes with significant difference (p=.05). I just know how to plot the power spectra of different variable comparisons with the option of certain p-values or "exact".
I've read the attached topic 'Get the p-value', but with the mentioned solution I don't get exact numeric p-values and the statistic test corresponding parameters of each electrode using power spectrum plots (alpha-frequency, beta etc.).

Kind regards,

Alexander

Alexander John

Trainings- und Bewegungswissenschaft
Raum: GH 113
Johannes Gutenberg-Universität
Albert Schweitzer-Straße 22
55099 Mainz
Email: alexjohn at uni-mainz.de<mailto:alexjohn at uni-mainz.de>
Tel: +496131 39 24560
Sprechstunde nach Vereinbarung

[Eeglablist] Get the p-value <https://sccn.ucsd.edu/pipermail/eeglablist/2016/011065.html>

Dear Edward and Andria,

I got two pieces of advice for you.

To obtain exact p-values for the significance-masked spectra,

1. The simplest way to do it is to select "exact" or empty in the p-value

edit box.

2. That should be done from the command window. The easy way should we

something like this (ie. for channel AF4):

[STUDY, specdata, allfreqs, pgroup, pcond, pinter] =

std_specplot(STUDY,ALLEEG,'channels',{'AF4'});

Looks like the solution 2 is the right way to do it with coding.

Makoto

On Mon, Apr 4, 2016 at 11:19 PM, Makoto Miyakoshi <mmiyakoshi at ucsd.edu<http://sccn.ucsd.edu/mailman/listinfo/eeglablist>>

wrote:

> Dear Edward,

>

> > I need to compare several channels to see if there is a significant

> difference between them, EEGLAB produces the statistical differences (in a

> Study) represented on plots.

>

> Do you mean that you have within-subject conditions you want to compare

> for multiple channels, or comparison across channels are your purpose?

>

> Makoto

>

> On Mon, Apr 4, 2016 at 7:51 PM, Makoto Miyakoshi <mmiyakoshi at ucsd.edu<http://sccn.ucsd.edu/mailman/listinfo/eeglablist>>

> wrote:

>

>> Dear Edward,

>>

>> Hmm I'm not exactly sure it will exactly meet your needs, but you may

>> want to try Simon-Shlomo Poil's NBT.

>> https://www.nbtwiki.net/

>>

>> Makoto

>>

>> On Fri, Mar 4, 2016 at 10:09 PM, Edward Wiskers <edwardwiskers at gmail.com<http://sccn.ucsd.edu/mailman/listinfo/eeglablist>>

>> wrote:

>>

>>> Hi all,

>>>

>>> I can apply the permutation test on the data (based on power spectrum)

>>> with having correction implemented, I need to compare several channels to

>>> see if there is a significant difference between them, EEGLAB produces the

>>> statistical differences (in a Study) represented on plots. However, I need

>>> to the p-value as a numeric number. Precisely, if I compare O1 and O2, I

>>> need to see the p-value of the permutation with correction for

>>> multi-comparisons as a numeric value, e.g., p=0.02.

>>>

>>> How to do that in EEGLAB

>>>

>>>

>>> Regards,

>>> Edward

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20171130/10c860af/attachment.html>
```