Dear colleagues,
We are running a two sessions protocol with 3 runs, each of them with 18 trials. A trial is composed by 2 sec of instruction presentation, followed by 2 sec presentation of a picture (6 neutral pictures and 12 negatives), then 4 sec of three equal balanced tasks and 4 sec of a fixation point. In the first level analysis, we are running a GLM with 6 regressors: INST, NEU, NEG, TASK1, TASK2, TASK3. The fixation point is the baseline. We are interested in the following contrasts that will enter in a second level group analysis:
NEG > NEU: [0 -1 1 0 0 0 ]
TASK1 > TASK2: [0 0 0 1 -1 0 ]
TASK2 > TASK3: [0 0 0 0 1 -1 ]
My question is as we have an unequal number of NEG (12) and NEU (6) trials it is correct to do the NEG > NEU contrast? If it is not, how I can fix that?
Also we have another issue concerning the second level analysis. For movement reasons, we excluded some subject's runs. So, for few subjects the number of runs in the first session is not the same of the second session. When doing a paired comparison between sessions, could I considere the averaged contrasts of runs for each subject?
Thank you for your advice,
Fernanda
|