Dear Mohd,
Based on a block length of 15 s and several different conditions it is likely that the default high-pass filter removes relevant parts of your signal / the low frequency part. Maybe you can provide some screenshorts of a representative design matrix (via "Review") after filtering?
> each condition had a lateralization cue, and each subject responded the lateral (left or right; by clicking a button) only once after the end of
> the block (and not for each TR).
Are you interested in these cueing effects? With a single predictor for left and right corresponding effects might cancel each other. In case you're especially interested in the motor response, it wouldn't be surprising if you failed to detect anything related as the motor execution takes place at the end of the block, thus a *block* regressor *starting much earlier* would be a very bad predictor.
Best
Helmut
|