An,
Yes, you can claim statistical significance for the cluster if that
cluster has a p value < 0.05. Note that you cannot say where, exactly,
in the cluster the activation happened. Only that it happened
'somewhere' in the cluster.
If you use FWE correction, then every active voxel is significant all
by itself, without regard to cluster size.
Jim
On 6/15/17, An <[log in to unmask]> wrote:
> Hi,
>
> i have two very quick and basic questions:
>
> is it good practice to threshold results at say p <.001 uncorrected and then
> look at the cluster-wise corrected statistic and then claim that a cluster
> is significant if p FWE < .05 ?
>
> is the FWE correction performed when clicking on the FWE button in the
> results tab voxel-wise, or it is still take into account the spatial
> structure of the signal ?
>
> Thank in advance for any input
>
|