Marek,
> Is there any spm method for reporting cluster level analysis in VBM?
If the assumption of stationarity isn't reasonable, no.
> I've read about SnPM, but we can not run matlab 4, so this is
> unavailable for us.
SnPM provides corrected p-values for cluster size without making an
assumption of stationarity.
> BTW, what is the last word on SnPM for SPM99?
While I am loath to quote an exact release date, we plan to have a
Matlab5 "SnPM99" out before Human Brain Mapping (2001).
-Tom
|