Dear Roberto,
I am trying to optimize my denoising per your recommended QC guidelines (Morfini et al., 2023) and am finding that the closest I can obtain towards the 95% threshold for QC-FC indices with my sample (using task-based connectivity with moderate ART settings w/ severe clinical populations [bipolar disorder +/- substance use disorder[s]) is if I apply polynomial expansions. Namely, with 48 parameters assigned to white matter and 32 parameters assigned to CSF I am able to achieve 93-94% for my QC-FC indices. I've tried everything and cannot hit the 95% threshold. Normalization QC, etc. looks great.
I attended one of your virtual workshops, which was awesome! In my notes, I jotted down that you do not recommend polynomial or temporal expansion for WM/CSF. Is my case any different? Is it okay to do so? At the moment I am shifting my pipepline from moderate to liberal ART settings to find another potential solution but would really appreciate your input either way. Thank you in advance!
Best,
Will