Table 5.

Classification performance of different structural graph metrics.

Method	Accuracy (%)	Sensitivity (%)	Specificity (%)	AUC
C	82.86	88.89	76.47	0.9027
G	55.71	61.11	50.00	0.6266
N	61.43	69.44	52.94	0.6756
MK_CG	84.29	88.89	79.41	0.9061
MK_CN	90.00	97.22	82.35	0.9509
MK_GN	62.85	72.22	52.94	0.6781
MK_CGN	91.43	97.22	85.29	0.9510

Structural connectivity (C), Global metric (G), Nodal metric (N); MK-SVM, multiple kernel support vector machine.