I used the same data for mean analysis with a threshold of 0.005 and the results were stable and consistent, but after correction by FWE, the results were not very consistent each time (the voxel range that comes with the system was not consistent each time)

for examlpe，once was “Blobs of ≥ 183 voxels”，the other was “Blobs of ≥ 234 voxels”

if any advice 
thanks so much

Ian

Dear Ian,

You can increase the number of permutations to achieve more stability.

With best wishes,

Joaquim