Skip to main content
. Author manuscript; available in PMC: 2007 Oct 1.
Published in final edited form as: IEEE Trans Biomed Eng. 2006 Oct;53(10):1883–1892. doi: 10.1109/TBME.2006.876620

TABLE 2.

The accuracy (in %) of identification by the IC method with the optimal penalty function C1 for (12) and C4 for (13). The EEG signals under the color noise with pre-whitening are used with source configuration of Case B (Fig. 1(b)): sinusoid sources with close frequency band (9.9 Hz, 10 Hz and 10.1 Hz). Three groups (ρ12, ρ23) are used, e.g., (0.5, 0.5), (0.5, 0.6) and (0.5, 0.7). ). In the simulations of (12), the noise information is accuracy. In the simulations of (13), two noise inaccuracy values are considered, e.g., κ = 1% and 7%. NL: Noise level.

Accuracy (%): (12) and C1
NL 5% 10% 20%  

One Dipole 99 99 99

ρ12, ρ23 0.5, 0.5 0.6, 0.5 0.7, 0.5

NL 5% 10% 20% 5% 10% 20% 5% 10% 20%

Two Dipoles 99 99 99 99 99 99 99 99 99
Three Dipoles 99 96 80 99 95 75 99 94 67
Accuracy (%) : (13), C4 and κ= 1%
NL 5% 10% 20%  

One Dipole 100 100 100  

ρ12, ρ23 0.5, 0.5 0.6, 0.5 0.7, 0.5

NL 5% 10% 20% 5% 10% 20% 5% 10% 20%

Two Dipoles 100 100 100 100 100 100 100 100 100
Three Dipoles 95 62 18 95 57 12 90 52 9
Accuracy (%): (13), C4 and κ = 7%
NL 5% 10% 20%  

One Dipole 100 100 100  

ρ12, ρ23 0.5, 0.5 0.6, 0.5 0.7, 0.5

NL 5% 10% 20% 5% 10% 20% 5% 10% 20%

Two Dipoles 100 100 100 100 100 100 100 100 100
Three Dipoles 95 68 31 94 66 26 89 57 18