Hi Jason
it depends on what correction you are referring to (in all cases this
not Bonferroni on clusters)
- for FWE the probabilities are nested from set to cluster to voxel, and
the correction is based on random field theory (knowing under H0 how
many blobs (the set) we have above a threshold, it is possible to derive
probabilities for clusters and voxels -- see random field theory stuff
on the MRC-CBU website, that's quite readable)
- for FDR that's different, after setting an excursion set (say
threshold at 0.001 uncorrected), the probability under H0 of finding
clusters of that size can be computed, and among those clusters, the
expected long run number of false positives is removed (see False
discovery rate on Wikipedia)
Cyril
> Dear SPMers,
>
> As I conceptually understand, the cluster-wise correction is to adjust
> the alpha level by accounting for a total number of clusters in the
> entire image (e.g., 0.05/500, if alpha is 0.05 and the number of
> clusters is 500).
>
> Am I right in assuming that the minimum unit of a cluster is two
> contiguous voxels, each of which gets passed a voxel-wise threshold?
>
> If anybody knows the mechanism of this closer-wise correction, please
> help.
>
> Jason
>
>
--
Dr Cyril Pernet,
Academic Fellow
Brain Research Imaging Center
http://www.bric.ed.ac.uk/
Division of Clinical Neurosciences
University of Edinburgh
Western General Hospital
Crewe Road
Edinburgh
EH4 2XU
Scotland, UK
[log in to unmask]
tel: +44(0)1315373661
http://www.sbirc.ed.ac.uk/LCL/
http://www.sbirc.ed.ac.uk/cyril
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.
|