Dear Rik:
sorry for the delayed response was busy with other stuff;
here are the results for the two designs:
Eff Contrast
Full_1_cue 1.4549 [1 0]
Full_1_video 3.6563 [0 1]
Diff 1.5855 [1 -1]
Full_3_cue 0.485 [1 0 1 0 1 0 0 0 0]
Full_1_video 1.2188 [0 1 0 1 0 1 0 0 0]
Diff 0.5285 [1 -1 1 -1 1 -1 0 0 0]
Red_1_Cue 1.5701 [1 0]
Red_1_Video 4.3286 [0 1]
Diff 1.5274 [1 -1]
Red_3_Cue 0.522 [1 0 1 0 1 0 0 0 0]
Red_3_Video 1.4381 [0 1 0 1 0 1 0 0 0]
Diff_Red_3 0.511 [1 -1 1 -1 1 -1 0 0 0]
Full means the design with two conditions (cue and video, 24 trials
in each); Red is the reduced 'design', where I skipped 7 of the video trials
following the cues to reduce regressor correlation; I attached the design
matrices.
As for the multi-session thing, you were right - when I concatenate the
sessions into one 'virtual session' efficiency goes up for all contrasts:
Concat_Cue 4.0685 [1 0]
Concat_Video 10.2352 [0 1]
Concat_Diff 4.7147 [1 -1]
Should any of this be worrisome, in any way?
TIA,
Claus
|