Table 7.

MCI/NC classification comparison with iMSF and Ingalhalikar’s algorithm. (Please refer to Table 6 for detailed descriptions of this Table.)

Framework	Methods	Performance measures				t-test on AUC (p-value)
		Acc	Sen	Spe	AUC	with rEM	with mFPC
Proposed	rEM	0.701 ± 0.047	0.866	0.414	0.774
	mFPC	0.715 ± 0.056	0.753	0.649	0.773
iMSF	regression	0.692 ± 0.063	0.649	0.768	0.760	0.001	0.001
	logistic	0.706 ± 0.051	0.818	0.509	0.733	< 0.0005*	< 0.0005*
Ingalhalikar’s	fusion1	0.642 ± 0.062	0.644	0.639	0.664	< 0.0005*	< 0.0005*
	fusion2	0.649 ± 0.063	0.651	0.643	0.689	< 0.0005*	< 0.0005*