Hi Alfonso,
Thanks so much for your help! Just a few follow-up questions:
1) I noticed that when I change the WM/CSF components in the denoising tab, this updates the QA plots. But I would still need to click "done" to actually run the denoising step with these new component numbers, right? Attached is what the new QA plots look like after increasing both the WM and CSF components from 5 to 15. It looks better, but still not great right?.... Would increasing them even more be too much? It looks like the one flat-looking subject/session in the plot starts to look like a bimodal pattern if I add 20... Maybe I can just remove that subject/session?
2) When I tried to change the outlier threshold in the Setup.Covariates tab, there was no option to "Compute new/derive first-level covariates". The only three options were: "Display covariate and single slice functional", "compute summary measures", and "reassign all covariate files simultaneously". Do I have an older version or something?
Is it better to try to increase the aCompCor components first, and then if the largest value still doesn't produce good QA plots, THEN to change the outlier threshold? Or try the outlier threshold if say 10 components for WM/CSF don't work, and then if that still doesn't produce good plots, try adding more aCompCor components and/or adding 2nd-order derivatives to realignment?
Thank you!
Kaitlin
Threaded View
Title | Author | Date |
---|---|---|
Kaitlin Cassady | Jul 10, 2023 | |
Kaitlin Cassady | Jul 21, 2023 | |
Alfonso Nieto-Castanon | Jul 26, 2023 | |
Kaitlin Cassady | Jul 26, 2023 | |
Alfonso Nieto-Castanon | Jul 27, 2023 | |
Kaitlin Cassady | Jul 29, 2023 | |
Kaitlin Cassady | Aug 9, 2023 | |
Alfonso Nieto-Castanon | Jul 20, 2023 | |