[Eeglablist] Help using LIMO BETA for STUDY

[Delorme, Arnaud]

Dear Nastassja,

> P.s.: I could figure out the problem. I just had to change the method from WLS to OLS and it ran the one-way ANOVA. However I have another question: I would like to perform a regression analysis with behavioural ratings in each trial as the continuous variable and the single-trial EEG data for multiple subjects. Unfortunately, it seems LIMO only accepts continuous variables in the single-subject level (e.g. it work if my continuous variable file would be the age of each participant, am I right?).

LIMO will accept continuous variables at both the single trial level (random effect level) and group level. At the group level, you have to select the regression option in the graphic interface and it will allow you to use a text file  containing your continuous variables with one value per subject.

Last year, there were restrictions on what you could do at the second group level (you could not mix continuous and categorical variables at the second level) but I think Cyril has fixed that since.

WLS (weighted least square) should work as well as OLS (ordinary least square) although there is a branch in LIMO named “WLS” that contains the latest WLS fixes. If one works and not the other, please share your dataset with us so we can look into it.

Best wishes,

Arno