> I just followed the steps and at the bottom of the table appears "FWEc: Inf". Do you know what this means?
There is no cluster which is sig. on cluster level.
> Should I leave it equal to zero?
When working with cluster correction it is irrelevant as you just loop through to obtain the FWEc, which is the one based on which you will reject clusters or not.
> Regarding the initial uncorrected threshold, should I stick to the 0.001 as suggested by Eklund et al. or in this case this doesn't apply?
I would go with the one that you use when considering whole-brain results. The SVC is really just meant to take into account the smaller volume, but if you also modify the initial threshold (e.g. liberalizing from .001 to .005) then you don't just correct for the smaller volume properties obviously. This is not incorrect, as you can come up with any arbitrary threshold (or combinations thereof) you like, but it's certainly somewhat inconsistent.
With regard to findings by Eklund et al., as far as I recall they did not look at the size of the volume, so I don't think one can draw any conclusions for SVC specifically.
Best
Helmut
|