Table 2.

STGC-GCAM model performance under different classification tasks. Values are reported in terms of mean ± standard deviation

CN vs. MCI					CN vs. AD				MCI vs. AD				sMCI vs. pMCI
Model	ACC	SEN	SPE	AUC	ACC	SEN	SPE	AUC	ACC	SEN	SPE	AUC	ACC	SEN	SPE	AUC
SVM	0.64 ± 0.003	0.52 ± 0.010	0.73 ± 0.010	0.62 ± 0.003	0.59 ± 0.004	0.32 ± 0.010	0.68 ± 0.004	0.50 ± 0.010	0.64 ± 0.004	0.21 ± 0.010	0.84 ± 0.010	0.53 ± 0.004	0.65 ± 0.010	0.30 ± 0.010	0.77 ± 0.010	0.53 ± 0.010
MLP	0.64 ± 0.004	0.53 ± 0.020	0.71 ± 0.010	0.62 ± 0.010	0.75 ± 0.003	0.22 ± 0.010	0.93 ± 0.010	0.58 ± 0.004	0.62 ± 0.004	0.14 ± 0.010	0.87 ± 0.010	0.50 ± 0.003	0.71 ± 0.010	0.21 ± 0.010	0.88 ± 0.010	0.54 ± 0.010
RF	0.63 ± 0.004	0.45 ± 0.004	0.76 ± 0.010	0.61 ± 0.004	0.74 ± 0.003	0.12 ± 0.010	0.96 ± 0.003	0.54 ± 0.004	0.69 ± 0.002	0.12 ± 0.010	0.94 ± 0.004	0.54 ± 0.002	0.64 ± 0.010	0.32 ± 0.020	0.74 ± 0.010	0.53 ± 0.010
LR	0.64 ± 0.003	0.48 ± 0.010	0.76 ± 0.010	0.62 ± 0.004	0.75 ± 0.003	0.21 ± 0.010	0.93 ± 0.002	0.57 ± 0.004	0.65 ± 0.010	0.11 ± 0.010	0.90 ± 0.010	0.51 ± 0.010	0.73 ± 0.010	0.14 ± 0.010	0.93 ± 0.010	0.53 ± 0.010
Brain CNN	0.72 ± 0.010	0.77 ± 0.010	0.65 ± 0.040	0.76 ± 0.010	0.76 ± 0.010	0.72 ± 0.010	0.78 ± 0.010	0.81 ± 0.010	0.64 ± 0.010	0.33 ± 0.040	0.79 ± 0.030	0.60 ± 0.004	0.77 ± 0.004	0.62 ± 0.010	0.82 ± 0.010	0.78 ± 0.010
GCN	0.83 ± 0.003	0.79 ± 0.020	0.85 ± 0.010	0.86 ± 0.010	0.81 ± 0.010	0.53 ± 0.050	0.91 ± 0.020	0.78 ± 0.030	0.68 ± 0.002	0.49 ± 0.010	0.77 ± 0.010	0.62 ± 0.010	0.79 ± 0.010	0.61 ± 0.010	0.85 ± 0.010	0.74 ± 0.010
STGC-GCAM	0.93 ± 0.001	0.87 ± 0.010	0.97 ± 0.003	0.98 ± 0.001	0.90 ± 0.002	0.99 ± 0.002	0.87 ± 0.002	0.95 ± 0.002	0.92 ± 0.002	0.97 ± 0.003	0.89 ± 0.002	0.96 ± 0.002	0.85 ± 0.002	0.67 ± 0.009	0.92 ± 0.006	0.79 ± 0.002

Abbreviations: ACC Accuracy, AUC Area under receiver operating characteristic curve, SEN Specificity, SPE Specificity