Dear Richard & SPM-experts,
similar to Paulīs example but assuming three quite orthogonal conditions
A, B & C - would you advice to explicitely model baseline for each
condition? I.e., modelling baselines seperately one would end up with at
least 7 columns (3 x [condition & baseline] + error term if you donīt
add temporal derivatives etc.) as opposed to 4. I guess my question is
whether:
1 0 0 (no explicit baseline modelling for A, B & C)
is really the same as
1 -1 0 0 0 0 (with explicit baseline modelling),
and more specifically if
-1 1 1 -1 0 0 would be computable to model Bīs activity at times where A
is not there and if this is indeed equivalent to
-1 1 0
TIA-
andreas
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|