help > Individual QC-FC measures are above recommended threshold but not when combined
Showing 1-3 of 3 posts
Display:
Results per page:
Jul 24, 2024  02:07 PM | Joseph Dust
Individual QC-FC measures are above recommended threshold but not when combined

Hi Alfonso and NITRC users,


I have a fairly nuanced question about the QA process that I've yet to find guidance on. I completed a QA report and confirmed that each of my main QC-FC association measures (MeanMotion, PVS, and Invalid Scans) each pass the recommended 95% threshold. However, the combined measures report below threshold at 93.8%. Is it recommended that both the individual and combined QC-FC associations meet the 95% or the individual measures sufficient. The subject-level QC measures do not show any extreme outliers >1.5 IQR for any of the substantial measures (invalid scans, PVS, mean motion, DOF). I have tried increasing the WM and CSF parameters to 15 parameters from 5 which results in the combined QC-FC measures to meet the threshold, however I'm not really sure if this is necessary or might remove important signal from dataset. I haven't really found any answers addressing this particular scenario, and I'm not sure why the combined QC-FC would be below the recommended threshold if the individual measures passed. Any guidance would be most helpful! 


 


Best,


Joe

Aug 14, 2025  06:08 AM | Sarah Valkenborghs - The University of Newcastle
RE: Individual QC-FC measures are above recommended threshold but not when combined

Hi Joseph, 


Did you ever find a solution to your query? Or how did you end up deciding to proceed? 


I'm in the same situation whereby my individual QC-FC measures are all >95%, but when I combine them it drops to ~89-90%, so I'm unsure what to do. 


I'd appreciate any pointers you might have, 


Thanks, 


Sarah V


 

Aug 28, 2025  10:08 PM | Alfonso Nieto-Castanon - Boston University
RE: Individual QC-FC measures are above recommended threshold but not when combined

Hi Sarah&Joseph


That is perfectly fine, in our recommendation for >95% percent overlap measures we were always referring to the individual measures, so as long as all of your individual percent overlap measures across the QC variables used in your QC-FC analyses are above 95% that should be sufficiently conservative to proceed (I will probably change the behavior of the "combined" QC-FC overlap measure to reflect this in the GUI, displaying instead just the minimum of the individual measures in order to avoid confusion). 


Hope this helps


Alfonso


Originally posted by Sarah Valkenborghs:



Hi Joseph, 


Did you ever find a solution to your query? Or how did you end up deciding to proceed? 


I'm in the same situation whereby my individual QC-FC measures are all >95%, but when I combine them it drops to ~89-90%, so I'm unsure what to do. 


I'd appreciate any pointers you might have, 


Thanks, 


Sarah V