Hi all-
Just curious here. 3dFWHMx is normally performed on errts files since it's the smoothness of the noise that matters. This makes perfect sense for fMRI tasks. However for resting state analyses the errts is "the good stuff" and not considered noise, per se. Should this matter for 3dFWHMx smoothness estimation? After normal resting state analyses the errts that result from a standard afni_proc.py stream are of course going to be fed into 3dTcorr1D, which doesn't output an errts itself. This makes me wonder if we should instead be using a single regressor in 3dDeconvolve for that 3dTcorr1D step, and then performing 3dFWHMx on
those errts files. A wild idea. In any case I'm pretty sure there's no statistical violation normally occurring so I'm mainly asking experts for their thoughts and some clarification. Thanks!
-Sam