Analysis of Residual Dependencies of Independent Components Extracted from fMRI Data

Table 2

Evaluation of ICA of the dataset with activations in ROI #1 and ROI #2 with time delay of 2.5 seconds at different noise levels and model orders. The AUC is shown near the IC index (in brackets).

Activations delay, 2.5 sec	Number of ICs	ROI #1; (IC index) AUC	ROI #2; (IC index) AUC
Noise; = 0.33%	5	() 0.91	() 0.89
	10	() 0.78 () 0.8	() 0.76 () 0.74; () 0.77
	15	() 0.8 () 0.8	() 0.8 () 0.8; () 0.71
	20	() 0.7 () 0.7; () 0.7 () 0.74	() 0.77 () 0.79; () 0.77
Noise; = 0.66%	5	() 0.82	() 0.78 () 0.79
	10	() 0.67 () 0.79; () 0.76	() 0.78 () 0.77
	15	() 0.77 () 0.78	() 0.78 () 0.8
	20	() 0.74 () 0.77; () 0.72	() 0.73 () 0.78; () 0.77
Noise; = 1.00%	5	() 0.71 () 0.77	() 0.82
	10	() 0.75 () 0.8 () 0.69	() 0.75 () 0.76
	15	() 0.73 () 0.7	() 0.7 () 0.74; () 0.72
	20	() 0.68 () 0.74	() 0.7 () 0.79
Noise; = 1.33%	5	() 0.75 () 0.67	() 0.68 () 0.75
	10	() 0.65 () 0. 71	() 0.76 () 0.61
	15	() 0.74 () 0.65	() 0.65 () 0.74
	20	() 0.74 () 0.64	() 0.69 () 0.67