Blistering Fast Functional Connectivity Analysis with 3dTproject

Because speed is associated with freedom, happiness, and the American way - we want results, we want visible results, and we want them now, darn it! - it is no surprise that neuroimaging analysis is becoming exponentially faster and more efficient. Gone are the days when you could run a slice-timing step, go to bed, and wake up eight hours later when it finally finished. Most likely it crashed, meaning you had to run the entire thing all over again, but that didn't matter - the point was that you at least felt as though you were doing something during that time.

Regardless, we are approaching a radically different era now; and one of the harbingers of that era is AFNI's 3dTproject. Released a few months ago, this tool is now the default in both uber_subject.py and afni_proc.py when you do resting state analyses. It's quicker, more efficient, and allows fewer chances to mess things up, which is a good thing.

To include 3dTproject in your analysis pipeline, simply apply "rest" to the analysis initialization screen when running uber_subject.py, or copy example #9 from the help documentation of afni_proc.py. Under no circumstances should you try running 3dTproject manually, since you, possessing the hand-eye coordination of a cheese burrito, will inevitably make a typo and screw something up. Nevertheless, if you insist on doing it yourself, I recommend using the following code that is automatically generated by the .py scripts:

3dTproject -polort 0 -input pb04.$subj.r*.blur+tlrc.HEAD -censor motion_${subj}_censor.1D -cenmode ZERO -ort X.nocensor.xmat.1D -prefix errts.${subj}.tproject

Needless to say, you will have to have already run your preprocessing steps and have run 3dDeconvolve with the -x1D_stop option to generate the necessary matrices.