Hi,
I am planning a new fMRI experiment. I'd like to know the best way to model the data, in order to answer my questions of interest, and to ensure that I don't start an experiment I can't use. It's a complex design, so I'm hoping that someone with experience can help here as this is not in the GLM cookbook or on the ML.
There are a couple of aspects to my planned study. The experiment will have a 2x2x2 repeated measures approach, where each factor has 2 levels. Here are the core questions I want to ask:
1) where is the BOLD signal modulated by the levels of each factor (i.e. the main effects)
2) where does the BOLD signal differ, as a function of these factors (i.e. all interactions; one I'm particularly interested in is where in the brain is factor 1 different to factor 2 and factor 3).
Lastly, I will run the same paradigm with patients and with controls (so will end up with a 2x2x2x2 mixed model, where the first factor is between-subjects. I want to see if there are differences between how my experimental manipulation modulates the BOLD signal. I am also expecting that patients will show a different pattern of laterality for one of the main effects. How best would I test these aspects?
Many thanks for advice/thoughts.
Best wishes,
GD
|