Basically, I am trying to move my lab away from the current script we use to detect spikes and I wanted to try out fsl_motion_outliers. I understand the basic functionality and I was able to play around using my data with some success. What I cannot figure out however, is whether I should be running the script on the pace or the moco data that I get from the scanner. My intuition is to run in it in the raw pace data and then use the generated matrix with my motion corrected (moco) data when running it in FEAT. Is this correct? I think running the fsl_motion_outlier script with the moco file would generate too many false positives.

I did notice however that the fsl_motion_outlier script has the option to add "--moco" as a parameter when generating the matrix and I am wondering if I should run my script using the moco data but with the "--moco" parameter and then use the moco file in the FEAT analysis. The third option would be to run it with the pace data and use the pace data when running FEAT? 

Which of these is the correct way to deal with this script.