Dear Manish,
>In one study we have apriori hypothesis and we do get significance
>in those regions after applying SVC but for some of them, the
>height threshold is as low as p=0.01, uncorrected, while for some of them
its
>p=0.0005, uncorrected. Also the box size that we select for SVC is not
fixed,
>for some of them its as high as 25*25*25 and for some of them as low as
>Do you think, its correct to use the different box dimensions and
>different thresholds if we have a hypothesis?
Yes, this is perfectly fine. You can have a series of
hypotheses that are more or less anatomically constrained
that require different volume SVCs. The differences in volume
will be expressed as differences in the uncorrected thresholds.
Critically, the corrected thresholds will all be the same
(i.e. p = 0.05).
I hope this helps - Karl
|