Hi Gang,
I want to find areas of activation once the Error has been removed.
1- I would certainly like to test for linear relationship. Would I have to add a factor and a column of [1 2 3 4 5 6 7 6 8] for each block?
2- But I'd also like to test without specifying any trend per se, because it's likely to change across brain areas and linear may not fit what's really happening. So I'd rather just look for areas with changes in at least 1 block; like an F test from a one-way anova would do, but with the Error removed.
What kind of Model would I have to write?
3- The model you proposed : Model:Error+Cond resulted in 4 subbriks. #0 Cond F, #1 Error F, #2 Error, #3 Error t. Is #0 the F test of the main effect of Cond once the Covariate has been removed or just the main effect of Cond? and #1 the main effect of Error (covariate)?
Thanks for your help,