It is more straightforward if you use -log10(p). If you want to do an
unsigned test, then run it twice, once with pos and once with neg, then
multiply the resulting cluster-wise p-values by 2.
doug
On 10/21/2014 02:11 AM, Joo Tan wrote:
>
> Hi
>
>
> I have a few questions regarding the use of “
Hi
I have a few questions regarding the use of “mri_volcluster” for
clusterwise thresholding. Essentially, we have a volume of pvalues
obtained from a two tailed t-test comparing beta values from condition A vs
condition B.
When I ran this command:
mri_volcluster --in transformedP.nii --sum