Dear SPM users,

 

I need your advice on how to analyze our complex fMRI study using SPM8. We have two groups of research volunteers (GroupA and GroupB). It’s a longitudinal study with 3 visits (B, 1Y and 2Y). And each visit has 3 scans with a visual attention task of varying difficulty (Easy, Normal and Hard).

As I understand, the appropriate design would be the Full factorial design. The first two factors seem obvious to me: status or group with 2 independent levels and visit with 3 dependent levels. Is that correct?

What I am confused about is whether it would be correct to include the task difficulty into this repeated measure ANOVA design as a 3^rd factor with 3 levels, dependent or independent. It is measurements from the same volunteer, and it is the same ball-tracking visual attention task, but the difficulty is different (tracking 2, 3 and 4 balls out of 10), so the expected response is different too. So my question is whether I should include it as a 3^rd factor with 3 dependent levels? Or maybe independent? Or should I create 3 group vs. visit designs for each task?

And the last question is whether I can include participants that don’t have all data for each visit and task for various reasons, like poor performance, bad motion, etc. Can I do this at all from a statistical point of view? Can SPM8 handle situations like this? Or should I include only those participants that have every single scan for every single visit?

Any ideas and suggestions are welcome. Thank you.

 

Regards,

Renat.