 |
 |
Dear SPM community, In my study, each participant performs the same task 3 times (5 conditions in the task). Normally, I would combine the s repeats at the first level in a fixed effects analysis, then run a RFX at the second level. However, I only have a pilot sample of 7 people. So, it's as if I need to do 2 FFX one after the other, first to combine sessions, then to combine my sample. But, I can't use con images in a second FFX. One possibility was to 'pretend' that the 21 data sets were in fact from 21 different people, and set up a huge FFX matrix. This runs, albeit with greyed out columns, and nothing showing in the matrix itself apart from the staggered constant columns at the right hand side. But, just because it runs, does this model make best statistical sense. What is the best way to approach this 'double' FFx situation? with grateful thanks Rachel ************************************************* Dr Rachel Mitchell Lecturer in Psychology, Durham University. Chair of Durham University NeuroImaging Collective (DUNiC). Facility Manager (Durham University) for their forthcoming MRI scanner. Governor, Tees Esk & Wear Valleys NHS Trust.