Dear SPM
I have two fMRI visual scanning protocols on two patient groups which I would like to combine together. The first is a partial brain sequence (TR 4s, 11 slices throught visual cortex, 120 scans) and the second is a wholebrain acquisition (TR 4s, 27 slices covering most of brain, 150 scans).
I want to combine the two to look at the response in the visual cortex but have a few questions about second level analysis. The study design is exactly the same for both except in the number of scans and involves a simple epoch based BABAB.. visual stimulation paradigm.
1) For RFX is there a general number of subjects which is considered too small for the analysis to be effective? I want to perform a separate RFX for the wholebrain group to look at responses outside the visual cortex but have only 8 patients and 11 controls. Would a fixed effects model be better?
2) Is is acceptable to include all the scans from both groups or should I cut out the last 30 scans for the wholebrain group. The design matrix, I understand, is supposed to be the same for all contrast images in RFX analysis but does this also mean number of scans?
3) Is there a way of entering covariates into the second level analysis model which are subject specific e.g. VEP latencies or markers of visual function.?
Thanks for your advice.
Ahmed
|