Denoising effects on functional connectivity estimates. (A) Additive effects of denoising sources on detection of seed connectivity in a group of 25 healthy participants studied during three sessions. GSR, global signal regression; ACC, anatomical CompCorr; HM, head motion estimates; OUT, head motion and global intensity outliers. Display threshold p < 0.001. Original data from NYU CSC TRT: subjects 1–25, sessions 1–3. (B) Denoising reduces structured noise in individuals. Left–Connectivity values histograms in a single healthy participant before (gray) and after (yellow) denoising including WM signal, CSF signal, estimated head motion, and outlier removal. Middle–Global signal variation before and after denoising. Right–carpet plot of voxel signal variation before (top) and after (bottom) denoising. Original data from NYU CSC TRT: subject 16, session 1. (C) Denoising increases sensitivity to, and specificity for, the language network. Effects of including denoising sources on detection of left inferior frontal gyrus seed connectivity are seen in a single participant. WM, white matter; CSF, cerebrospinal fluid; HM, head motion estimates; Outliers, head motion and global intensity outliers. Display threshold r = 0.4. Original data from NYCSC TRT: subject 16, session 1.