Hi,
I have longitudinal data for an fMRI task (one run at time 1 and one run at time 2) and I was wondering if there is a difference between running a paired t-test at the group level and a 3 level analysis where I calculate the difference between the time 1 and time 2 scans for each subject individually (using fixed effects modelling) and then run a group (mixed effects) analysis on those difference images?