Dear all,
Happy New Year!
I am using SPM8. I am trying to simulate some task evoked time courses in a square of 64x64 voxels. I artificially made 4x4 voxels have task activation and the remaining voxels just full of Gaussian noise. The reason I do this is to test my processing procedure. Then I use SPM to run GLM, T statistics on Beta1(paradigm regressor) and looked at result for T statistics. I used FWE, no mask and no cluster. I see the graphic window display my simulated square on top of the brain template, and the threshold T value is around 4.10 which is the value that I typically got using true brain volumes. I am wondering how SPM calculate the FWE T threshold. Does SPM always use a template brain to calculate T threshold or calculate based on the input data that might be simulated square brain or subject individual brain?
Thanks,
Susan
|