Hello Rik Henson and/or other SPM'rs that have time and knowledge to share an answer,
I want to control for (between condition) reaction time in a 2x2 factorial design.
I collapsed factors as parametric modulations, and add rt as paramatric modulator to
control for global rt effects according to:
https://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind06&L=SPM&P=R750063
I can create a regressor for all trials (factor A + B), and add each factor as a parametric
modulator by:
1) using 2 modulations
1) factor A: A1 as [1 1 1 1 ...] and A2 as [-1 -1 -1 -1 ...]
2) factor B: B1 as [1 1 1 1 ...] and B2 as [-1 -1 -1 -1 ...]
or
2) using 4 modulations
1) A1 as [ 1 1 0 0 0 0 0 0]
2) A2 as [ 0 0 1 1 0 0 0 0]
3) B1 as [ 0 0 0 0 1 1 0 0]
4) B2 as [ 0 0 0 0 0 0 1 1]
I prefer the second option because I want to take effects of each seperate condition (A1,
A2, B1, B2) to a second level factorial design.
Obviously, this won't work with option 1. However, using option 2 gives some problems in
the SPM design: one of the parametric modulators are zeros (i.e. black in the design).
I guess this is due to the redundancy Rik is talking about in the thread mentioned above.
I tried 3 param mod's (N-1), but I'm not sure which contrasts to use to capture the effects
for each condition: Maybe [0 1 0 0 0], [0 0 1 0 0 ], [0 0 0 1 0 ] and [1 -1/4 -1/4 -1/4 0] ?
These effects doesn't look quite the same as in the original categorical design though.
In short: Can somebody help me to get beta's for each seperate condition (A1, A2, B1,
B2) in a parametrical design, reflecting the same effects as if I used a categorical design
with A1, A2, B1, B2 as seperate regressors?
Thanks,
Martijn
|