Hi Everyone,
I have a question on computing betas related to the following thread:
[
afni.nimh.nih.gov]
I have an experiment with 4 runs. In the past, I have fed each run into 3dDeconvolve (see below) and get an overall beta for each experimental condition of interest which I then use in group analyses. However, I now understand that it is best to compute beta coefficients seperately for each run and then either average them or use RUN as a within subjects variable in an ANOVA analysis so as to be able to look at order/time effects. My question is: Is the overall beta the average of all four betas (for each of the 4 runs) or is based on what would effectively be 1 long run?
Much thanks as always for your help.
Jatin