I have two questions that are both somewhat related to the AlphaSim program.

1. In browsing the AFNI website I have come across the reasonable heuristic that the fwhm used in the Gaussian blurring of our orig data would be a good level to set our Gaussian filter for the AlphaSim program.
We had used a blur distance of 3 for our orig data, but recently found that the output of 3dFWHM actually gives us parameters that look more like x 5 y 4 z 4.
Is it more appropriate to use a Gaussian blur level representative of the output from 3dFWHM in the program 3dmerge blurring of our orig data sets (about 4 or 5), or is the seemingly conservative 3 we had been using just as appropriate?

2. I found some reference to this topic, but could not discern a definitive answer.
Is it appropriate to use a mask with a truncated number of voxels, based on a t threshold that shows moderately large areas of activation, as the input for the AlphaSim program when we are not doing an eventual ROI analysis on our data, and if so than should we make a mask that “looks good” as said input, or is there a more precise way?

I realize the root of my problem may come from the lack of inherent guidelines for this type of research, but would still greatly appreciate any feedback.
Thank you so much for your time!
Jeremy