 |
 |
Hi, I have a question regarding how to include the baseline=rest periods into the FEAT full model design. We are analyzing data from a parametric static force experiment, trying to find the common rest-force response and any parametric force response. We have 5 different sorces: 200g, 400g, 600g, 800g, 1000g; I use 10, 20, 30, 40, 50 to represent them in the design respectively. The 5 different forces are each applied once per run in a pseudorandom order for two runs. Example : rest - 30 - rest - 10 - rest - 40 - rest - 20 - rest -50 - rest Just focusing on a single run, in the EV setup, I use one entry per column. I decided to use 6 original EVs so we could study the contrast between each force and rest (baseline), and as a function (linear or nonlinear) of force. In particular I want to look for parametric force effects AFTER I have removed any common ON-OFF force effects. My EV setup is like this: EV-baseline: 1-0-1-0-1-0-1-0-1-0-1 EV-30: 0-30-0-0-0-0-0-0-0-0-0 EV-10: 0-0-0-10-0-0-0-0-0-0-0 EV-40: 0-0-0-0-0-40-0-0-0-0-0 EV-20: 0-0-0-0-0-0-0-20-0-0-0 EV-50: 0-0-0-0-0-0-0-0-0-50-0 After i created a text file for each EV and input it into the GLM design, i got the rank deficiency error. If i don't include the EV-baseline, it works, and the EV-baseline by itself works. So it seems I can test for either an ON-OFF force effect or a parametric effect but not both simultaneously. I tried to find the answer in previous posts, i found in one reply you said '...Having said that, it's probably best *not* to include rest *explicitly* in the model because you will make it rank deficient...' Therefore, is it true to say that in order to avoid having to deal with rank deficient designs you do not allow the simultaneous testing we are trying to set up? thanks! Heather