 |
 |
Try this: http://fmripower.org/ And see this: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3427872/ Power in fMRI is a tricky business. Like any power analysis you need to have certain inputs, such as expected effect size and variance. But these remain largely unknown and are far more complicated in an fMRI context than in a behavioral design. It's tough that granting agencies want this so much, because any answer you give them will be somewhat contrived (we guess some parameters), or in the above case use data from a previous study which may be more or less valid when considering our future study). But the above software can give a good start point if you can make use of it. For my part I always also try to refer to other studies with similar types of designs which found significant results at appropriately controlled threshold (e.g. no 5 voxels p > 0.005 uncorrected, but FDR corrected significance, for example). And then if possible show you will collect an even larger sample if possible, so you should have power to detect relevant effects. Good luck! Colin Hawco, PhD Neuranalysis Consulting Neuroimaging analysis and consultation www.neuranalysis.com [log in to unmask] -----Original Message----- From: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]] On Behalf Of Mike Sent: November-27-15 5:29 AM To: [log in to unmask] Subject: [SPM] calculation of power Hi SPM experts, I design an event-related fMRI paradigm but I wonder how many repetitions of a task and how many subjects I should collect to reach sufficient power in future data analysis? Any methods to calculate these information? Thanks. Mike