[Eeglablist] Statistics for ERSPs

[Laurent Sheybani]

Dear all,

I am trying to obtain a topographical representation of t- and/or p-values of ERSPs using eeglab's STUDY.

My problem is that while I can obtain the average time-frequency representation on scalp topographies, I do not find the way of obtaining the statistical values.

Here's my procedure:
I preprocessed the data (re-referenced, downsampled, notched, baseline corrected them) and then 'epoched' them based on specific triggers (let's call them 'pos', 'neg' and 'ctrl'). Each epoch lasts 4000 ms. My goal is to see whether there is a difference in a specific frequency band (e.g., 8-12 Hz) during a certain time window of activity centered on the triggers 'pos' vs 'ctrl' and 'neg' vs 'ctrl'. I would like to get the topographies of p-values and t-values of the pre-specified frequency band during this given time window.

Thus, I believe I have one subject, 3 conditions ('pos', 'neg' and 'ctrl') and 1 group (1 group, because I want to apply paired statistics - is this the right way?).

In 'Edit STUDY design', I have 3 categorical variables that I do not combine together.

I prepared the channels including ERSP, but when I plot the results using the next window, I do not obtain neither t- nor p-values topographies. I only get the average topographies of frequency activity.

Would you have any idea of where I am missing a step to get the statistics of the ERSP topographies ?

The same procedure does not work neither with ERPs.
I use EEGlab version 2019.

In short, I do not obtain the 4th plot that is presented on this video, at 17:07:
https://youtu.be/HAGrNeh-kgA


Thanks in advance for your help !
Laurent Sheybani