> I run the group analysis, I get some negative correlation result in one region
> that show positive activity during the A condition, not duirng the B condition. I
> don't know how to interpret this negative result.

Negative correlation means that increased activity in the seed region would decrease the contrast between A and B.

> I did add the seed time series into the 3dDeconvolve as one regressor, and
> then I got beta value for this seed sequence. If I used this beta value to do
> group analysis, what is result means?

Supposedly that beta reflects the correlation between the seed region and the rest of the brain.

> based on the way I used, the seed time series is output from the '3dmaskave'.
> This time series includes the effect from the both A and B condition. Does this
> affect the pure interaction between the seed region and conditions contrast
> effct? Should I use the '3dSynthesize' to remove the baseline and conditions' effect?

No, you don't need to remove the effect from baseline and other conditions because they are actually zeroed out when you created the contrast of A and B with 1, 0, and -1's.

Gang