Skip to main content
. 2014 Apr 18;9(4):e95493. doi: 10.1371/journal.pone.0095493

Table 2. “Signatures” of Component Groups.

All Noise (M1) Eyeballs (M2) Head Motion (M3) Ventricles (M4)
1 Percent total activation in GM Percent total activation in eyeballs Percent total activation skull Percent total activation ventricles
2 Kurtosis of IC distribution Percent total activation in GM Percent total activation in WM Percent total activation in WM
3 Percent total activation MNI152all edges Spatial Entropy of IC distribution avg jump max min Caudate R
4 Percent total activation in CSF Avg distance btw 10 local max Percentage activation voxels LR symmetric Percent total activation in GM
5 Angular L Max cluster size 10 local max regiongrowing thresh 2.5 Amygdala R Caudate L
6 Angular R Mean cluster size 10 local max regiongrowing thresh 2.5 Percent total activation MNI152 all edges
7 Cingulum Ant R Percent total activation in WM power band 0.008 to 0.02 Hz
8 Temporal Sup L Skewness of IC distribution Percent total activation in CSF
9 Avg distance btw 10 local max Percentage activation voxels LR symmetric four lag auto correlation
10 Temporal Sup R Percentage activation voxels LR symmetric Amygdala L
Weights
1 0.5586 0.5032 0.549 0.5201
2 0.2665 −0.1894 −0.1565 0.2634
3 −0.2467 0.1884 −0.1077 0.2459
4 −0.2153 0.1797 0.0893 −0.2025
5 0.2098 −0.0674 −0.0711 −0.0674
6 0.1924 −0.0653 0.0571
7 0.1859 −0.0444 0.0519
8 0.1637 0.0354 0.051
9 0.1619 0.0288 0.044
10 0.1604 −0.017 −0.0425

Top 10 Chosen Features for classifiers M1, M2, M3, and M4 built using Data A (top), and associated weights (bottom). Gray Matter (GM), white matter (WM), cerebral spinal fluid (CSF). complete set of results included as Tables S1.