Dear Ines,
You can't compare the results of two different one-sample t-tests this way. With a one-sample t-test you test whether the beta values differ significantly from zero or not. If the variance is high in one group, then it might well be that you don't find any significant voxels in the corresponding one-sample t-test (= "no activation"), although the beta values might be higher on average compared to another group with rather low beta values but little variance (which might lead to highly significant voxels = "lots of activation").
Second aspect, the cluster size. You can have large significant clusters with rather small and/or inconsistent beta values compared to small clusters with large and/or consistent beta values.
In short,
1) high t-values do not mean high beta values (or high PSC)
2) large clusters = widespread activations do not mean high beta values (or high PSC)
Best,
Helmut
|