Dear Khalaf,
> Dear SPMers, I have a question regarding Friston's technical note (Testing
> for anatomically specified regional effects in Hum Brain Map :5: 1997):When
> testing for anatomically prespecified regions, does the p-value assigned to
> the activation spatial extent (nearest to hypothesized region) correspond
> to the cluster level p-value or the uncorrected p-value for spatial extent
> (as shown in SPM96).
It corresponds to the uncorrected p-value for spatial extent. This is
the probability of any given containing the observed number of voxels
or more (conditional on that cluster existing). Given that the cluster
has been selected, arbitrarily, as the nearest to some point this
probability is a p value for the regional effect in question.
With best wishes - Karl
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|