Dear experts,

For SDM-PSI, the default number of permutations is set to 1000 to get the null distribution, while in articles using SDM-PSI often 5000 is used (5000 permutations are also used for non parametric testing in FSL I believe) - are 1000 permutations sufficient to get reproducible results or would you advise to use 5000?

Thanks in advance, also for developing such great free software!

Dear Jorge, 

You can set the number of permutation depending of the precision you want to obtain the results. The maximum p-value that you can obtain with perm=1000 would be 1-1/1000 (0.999), whereas with perm=5000, would bbe 1-1/5000 (0.9998). Th default parameter is usally good enough, but if you want to obtain more precise results, you should increase the number of permutations.

Kind regards, 
Lydia

Moltes gracies Lydia. Best, Jurriaan