Hi,
In the experimental condition, after a 2 s visual cue, I delivered a 20 ms tactile stimulation (TS-E; E = experimental) to my subject. After the stimulus terminated, subject had to maintain a 5980 ms working memory (wm) interval (WM-E), which was followed by a response cue. The control condition have the same structure (i.e., with EVs of TS-C and WM-C; C = control) but subjects did not have to perform wm. I'm interested in activity related to stimulation itself and wm-related activity.
<Solution 1>
In the first first-level GLM (GLM-1), I created 4 EVs to obtain activations related to stimulation and wm:
For stimulation, the EV was with a duration of 0.02 s and an onset at the start of tactile stimulus.
For wm, the EV was with a duration of 5.98 s and an onset at the start of wm period (i.e., at the offset of tactile stimulation).
Thus, I had 4 EVs: TS-E, TS-C, WM-E, WM-C.
For stimulation-related activity, my contrast is "0 1 0 0" (TS-E was not included since it might contain some wm activity).
For wm-related activity, my contrast is "0 0 1 -1."
However, I wonder if it's possible to separate fMRI signal of wm from that of stimulation?
<Solution 2>
Therefore, I adopt another strategy. I built a second GLM (GLM-2) with only 2 EVs: E and C. Each of them was with a duration of 6 s and an onset at the start of stimulation.
For stimulation-related activity, I used above-described GLM-1 and the contrast "0 1 0 0".
But for wm-related activity, I used GLM-2 and the contrast "1 -1".
Does Solution 2 make more sense?
Thanks. Mike
|