Hi,
I would like to get your opinion about Montecarlo statistics. I know that there is maybe not a simple answer and that it depends probably on the research question but anyway, is there some kind of gold standard about the number of montecarlo simulations / p values ?
Speaking about that with colleagues, I had plenty of divergent opinions for both variables depending on their neuroimaging field i.e minimum 10,000 / 3,000 / 1,000 simulations - minimum p < 0.05 / 0.001 (the latest threshold being advised on the basis on the recent papers on false positives) .
I am also wondering if it makes sense to use this correction as a "SPM-like small volume correction".
What I mean by SPM-like small volume correction, is when you got clusters in a region for which you have an apriori hypothesis, that they do not survive to multiple comparisons, and you restrain your analysis to the region of interest. In my study, it would be 1 or 2 merged labels from the auditory cortex.
Thank you for your input,
Best,
Damien