Hi,
Sorry for a simple question but I really hope some experts to clarify my question.
In my experiment, subjects see a cue period indicating task type (2 types of task: t1,t2), which is followed by a stimulation period (2 types of stimulation: s1,s2). t1 is always followed by s1. 50% of t2 is followed by s1, and another 50% of t2 is followed by s2. The first-level GLM contains 2 regresses for cue period (t1, t2) and 3 regressors for stimulation period (t1s1, t2s1, and t2s2).
If my aim is to compare activation during stimulation period between t1s1 and t2s1, is there a multi-collinearity problem (because t1s1 always follows t1, so activity during t1s1 is contaminated by t1)?
Best wishes,
Manuel
|