Dear Helmut,
I figured out the issue: instead of declaring the second timepoint scans as a new session, I declared them as the same session as first timepoint but as different conditions. I still cannot explain why but after fixing the model, I no longer get the bizarre activations.
Back to the original post on contrasts and after reading Szameitat et al. 2011, is the right approach
To check for activation:
1) run sum of single tasks contrast LR - L – R aka [ 1 -1 -1 0] to check for activation
2) If there is activation, use results and conclude LR specific activation as this is the most conservative contrast. Need not run the interaction and conjunction contrasts to check for activation
3) If no activation, run the interaction contrast (LR + REST) – (L + R) aka [1 -1 -1 1] to check for activation
4) If there is activation in interaction contrast but not in 1), need not run the conjunction contrast to check for activation. How do we interpret this activation finding?
5) If no activation, run the conjunction contrast (LR – L) ∩ (LR – R) aka select both [1 -1 0 0] [1 0 -1 0] to check for activation
6) If there is activation in conjunction contrast but not in 1) and 3), how do we interpret this activation finding?
7) If no activation in all 3 contrasts, conclude no detectable LR specific activation.
To check for de-activation:
1) run sum of single tasks contrast -LR + L + R aka [ -1 1 1 0] to check for de-activation
2) If no de-activation, conclude no LR specific de-activation as this contrast is the most liberal for underadditive effects. No need to run interaction or conjunction contrast to check for for de-activation
3) If there is de-activation, run the interaction contrast (LR + REST) – (L + R) aka [-1 1 1 -1] to check for de-activation
4) If there is de-activation in interaction contrast and in 1), take the results from interaction contrast and need not run the conjunction contrast to check for de-activation. How do we interpret this de-activation finding?
5) If no de-activation in 1) and 3), run the conjunction contrast (LR – L) ∩ (LR – R) aka select both [-1 1 0 0] [-1 0 1 0] to check for de-activation
6) If there is de-activation in conjunction contrast but not in 1) and 3), how do we interpret this de-activation finding?
I came across the terms under-additive, additive and over-additive activation. I'm thinking activation refers to over-additive activation and de-activation refers to under-additive activation. What is additive activation?
Thank you very much!
Pei Ling
|