[FieldTrip] simple main effects and permutation

[Xinyi Li]

Hi all,


I have a mixed design and my hypothesis is about simple main effects. So for example, I have two groups of people, and each person experience the same 2x2 factorial design with conditions A1, A2, B1, B2. And my hypotheses are something like the simple main effect of A within group 1 and condition B1.


My questions:

1) Can I just subset the data to include only the group and condition I want (B1 & group 1) and do a t-test between condition A1 & A2 after subsetting? If I understand correctly, this approach will bias the standard error of the estimates? But I'm not sure if this matters in a permutation framework, and I also don't know if it's a common practice to do this in EEG analysis?

2) Alternatively, I can run a full mixed ANOVA model and then the simple main effect in R to get the test statistics I want. But for this approach I'm not sure how I should perform the permutation since fieldtrip doesn't support a mixed design. Should I only permute A1 & A2 within condition B1 and group 1? Or should I permute everything within both factors A & B? And what about the group labels?


Any suggestions? Thanks!


Xinyi
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20170919/bcbc8202/attachment-0001.html>