Hi FSL'ers,
Hope you can help me with a question about including a first level covariate of interest, and whether it is necessary to demean this value..
I am interested in the added effect of the 'value' of my current stimulus (besides the decision-processes, which is as long as the reaction time)
In my current model there are 2 EV's with a 3 column format:
first column displays the onset, second the duration (=RT), and the third the value of the stimulus.
In the first EV the 3rd column (value of the stim) is all 1's
In the second EV the 3rd column (value of the stim) is included
So 'visually' my current model looks like this:
EV1 EV2
1.5 2.5 1 1.5 2.5 10
2.4 3.5 1 2.4 3.5 50
3.6 2.5 1 3.6 2.5 40
Contrast: -1 1
Two questions:
1. I'm orthogonalizing EV2 with respect to EV1 - Is the current model, including the orthogonalization, correct?
2. Do I need to demean the value in EV2 (in the 3rd column)? - My stimuli values range from 4 to 150 - so demeaning it would lead to negative values as well.
What to do?
Hope you can help, thanks in advance!
Anna
|