Dear SPM List,
My understanding from searching the archives is that the cluster level
correction in SPM2 (5?) is not as accurate for smaller search volumes.
Am I correct in this belief? If that is indeed the case, is there a way
to get an upper bound on the corrected p value based on the search
volume size? If it is the case that the correction cannot be trusted at
small search volumes but an upper bound cannot be determined, is there a
heuristic for a volume size that would be safe? 1000mm^3? 5000mm^3? Is
there an interaction with the voxel threshold used?
This is all assuming the correction is done on an anatomic region of
interest based on an atlas such as the WFU atlas or the probabilistic atlas.
Thank you in advance for any information
Jason
--
Jason F. Smith, _.--~~~.~-._
Brain Imaging and Modeling Section .( ./ ).
NIDCD/NIH ( { ).
and (' / _ ./ )
Department of Psychology (' .~~~' / .}
Arizona State University ( .~~ )
[log in to unmask] '~~( ,'(((()))'
[log in to unmask] '..:~~' {{((())
{{}}
Rm 8S235B, 10 Center Dr. Bethesda MD 20892-1407
Ph) 301.451.1647 Fx 301.4805625
|