Table 4.

A summary of the performances for MCI vs. CN classification

Method	Modality	ACC (%)	SEN (%)	SPEC (%)	BAC (%)	PPV (%)	NPV (%)
SMTL	MRI	70.90 ± 3.04	80.78	52.00	66.39	76.30	58.58
	PET	64.98 ± 2.68	75.89	44.00	59.94	72.27	48.69
	MP	66.76 ± 3.29	73.56	53.67	63.61	75.35	51.33
	MPC	68.32 ± 3.48	74.89	56.00	65.44	76.14	54.32
DW-SMTL	MRI	68.89 ± 2.85	76.67	54.00	65.33	76.13	54.74
	PET	64.31 ± 2.82	73.89	46.00	59.94	72.34	47.96
	MP	70.94 ± 3.04	80.89	52.00	66.44	76.23	58.84
	MPC	72.77 ± 3.40	79.78	59.33	69.56	79.00	60.48
S2MTL	MRI	70.32 ± 3.18	82.78	46.67	64.72	74.66	58.81
	PET	67.60 ± 3.22	78.89	46.33	62.61	73.47	53.81
	MP	69.65 ± 2.56	76.78	56.33	66.56	76.66	56.49
	MPC	67.02 ± 2.95	78.78	44.67	61.72	73.07	52.55
D-S2MTL	MRI	68.85 ± 3.15	76.56	54.33	65.44	75.94	55.17
	PET	68.89 ± 2.96	76.89	54.00	65.44	75.68	55.66
	MP	70.98 ± 2.91	77.78	58.33	68.06	77.64	58.53
	MPC	68.98 ± 3.30	76.78	54.33	65.56	75.95	55.47
DW-S2MTL	MRI	77.57 ± 2.92	90.89	52.00	71.44	78.42	74.83
	PET	74.90 ± 2.55	96.00	34.67	65.33	73.69	81.97
	MP	80.11 ± 2.64	93.89	53.67	73.78	79.54	82.07
	MPC	78.77 ± 2.47	90.78	56.00	73.39	79.64	76.21

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

SMTL sparse multi-task learning, S ²MTL subclass-based SMTL, DW-SMTL deep weighted SMTL, D-S ²MTL deep S²MTL, DW-S ²MTL deep weighted S²MTL