 |
 |
How did this 11 come out? You need at least 20 to be able to obtain results with p = 0.05 (i.e. 1 out of 20). But this is only a theoretical limit. Usually, you need much more - min. 5000 is recommended. Higher number increases the resolution of the "histogram under null". I generally use 10000. Auer, Tibor M.D. Ph.D. Biomedizinische NMR Forschungs GmbH am Max-Planck Institut für Biophysikalische Chemie Am Fassberg 11 37077 Göttingen Germany Phone/Work: +49-(0)551-201-1725 Phone/Home: +49-(0)551-387-0076 Mobile: +49-(0)176-8012-7921 Mail: [log in to unmask] > -----Original Message----- > From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On > Behalf Of Maria Tziraki > Sent: Tuesday, August 21, 2012 12:50 PM > To: [log in to unmask] > Subject: Re: [FSL] randomise slow to run > > Hello, > > Thank you for your suggestions. May I ask you one last thing, the number of > permutations per contrast seems to be 11, is this correct? Should I run > randomise with more than 10000 permutations because of the model I am > using? Thank you in advance. > > Maria