Table 3.

A summary of the performances for AD vs. CN classification

Method	Modality	ACC (%)	SEN (%)	SPEC (%)	BAC (%)	PPV (%)	NPV (%)
SMTL	MRI	86.55 ± 3.02	88.33	84.33	86.33	87.55	85.28
	PET	80.45 ± 3.26	80.33	80.67	80.50	88.49	69.11
	MP	87.64 ± 3.31	90.33	85.00	87.67	85.53	89.96
	MPC	92.45 ± 2.69	94.00	90.67	92.33	92.04	92.94
DW-SMTL	MRI	88.36 ± 3.20	84.33	92.33	88.33	91.55	85.68
	PET	82.45 ± 3.22	82.33	82.67	82.50	89.70	71.85
	MP	90.45 ± 3.13	90.33	90.33	90.33	94.67	83.65
	MPC	92.45 ± 2.69	94.00	90.67	92.33	92.05	92.94
S2MTL	MRI	86.55 ± 3.02	78.33	94.33	86.33	92.90	82.14
	PET	85.36 ± 3.29	84.00	86.67	85.33	85.86	84.90
	MP	93.18 ± 2.52	90.00	96.33	93.17	96.05	90.68
	MPC	92.36 ± 2.70	94.00	90.33	92.17	92.33	92.40
D-S2MTL	MRI	83.64 ± 3.70	84.33	83.00	83.67	89.78	74.48
	PET	89.36 ± 2.68	90.00	88.67	89.33	89.54	89.16
	MP	88.27 ± 2.78	82.00	94.33	88.17	93.32	84.43
	MPC	90.36 ± 2.77	90.00	90.33	90.17	89.25	91.33
DW-S2MTL	MRI	90.36 ± 2.53	82.33	98.33	90.33	98.00	84.86
	PET	89.27 ± 2.97	82.33	96.33	89.33	95.80	84.27
	MP	93.18 ± 2.82	90.00	96.33	93.17	96.05	90.68
	MPC	95.09 ± 2.28	92.00	98.00	95.00	97.74	92.86

Boldface denotes the best performance and the maximum performance in each metric

SMTL sparse multi-task learning, S² MTL subclass-based SMTL, DW-SMTL deep weighted SMTL, D-S²MTL deep S²MTL, DW-S²MTL deep weighted S²MTL