help > QC-FC associations
Showing 1-1 of 1 posts
Sep 29, 2022 02:09 PM | iman_idrees
QC-FC associations
Hi all,
After running the denoising step in CONN, I had a look at the QC plots, specifically the QC-FC associations, to evaluate the quality of the denoising procedure. Interestingly, I found that before denoising, the data was closer to the null-hypothesis distribution with a match of 99.5% (for max motion) relative to 89% following denoising. By modifying the number of WM components and realignment/CSF parameters, I have increased the match slightly, however, the discrepancy between the figures before and after denoising remain in several measures (mean motion, condition specific GCOR measures). In this case, would you suggest that I do not denoise the data? Or perhaps add additional denoising steps?
Many thanks in advance,
Iman
After running the denoising step in CONN, I had a look at the QC plots, specifically the QC-FC associations, to evaluate the quality of the denoising procedure. Interestingly, I found that before denoising, the data was closer to the null-hypothesis distribution with a match of 99.5% (for max motion) relative to 89% following denoising. By modifying the number of WM components and realignment/CSF parameters, I have increased the match slightly, however, the discrepancy between the figures before and after denoising remain in several measures (mean motion, condition specific GCOR measures). In this case, would you suggest that I do not denoise the data? Or perhaps add additional denoising steps?
Many thanks in advance,
Iman