Table 2.

Logistic regression to predict amyloid-beta positivity in patients with mild cognitive impairment using demographics and magnetic resonance image features.

	Logistic regression			Support vector model			Random forest
	Mean	SD	p	Mean	SD	p	Mean	SD	p
Step 1 Demographic features
Age, sex, education	0.52	0.08	< 0.001	0.51	0.08	< 0.001	0.55	0.08	< 0.001
Age, sex, education + MMSE (total)	0.58	0.09	< 0.001	0.56	0.09	< 0.001	0.64	0.08	0.059
Age, sex, education + SNSB (detailed)	0.69	0.08	Ref	0.69	0.08	Ref	0.66	0.08	Ref
Step 2 T1-weighted magnetic resonance image
Age, sex, education + SNSB (detailed)	0.69	0.08	Ref	0.69	0.08	Ref	0.66	0.08	Ref
Age, sex, education + SNSB (detailed) + T1b (thickness)	0.68	0.09	0.03	0.67	0.09	0.008	0.62	0.09	< 0.001
Age, sex, education + SNSB (detailed) + T1 (texture)	0.71	0.08	0.06	0.71	0.08	0.12	0.70	0.08	< 0.001
Age, sex, education + SNSB (detailed) + T1 (volume)	0.73	0.08	< 0.001	0.73	0.08	< 0.001	0.73	0.08	< 0.001
Step 3 Diffusion-tensor magnetic resonance image
Age, sex, education + SNSB (detailed) + T1 (volume)	0.73	0.08	Ref	0.73	0.08	Ref	0.73	0.08	Ref
Age, sex, education + SNSB (detailed) + T1 (volume) + DTI (FA)	0.73	0.08	0.60	0.73	0.08	0.42	0.73	0.08	0.85
Age, sex, education + SNSB (detailed) + T1 (volume) + DTI (MD)	0.74	0.08	0.20	0.73	0.08	0.20	0.73	0.08	0.64

SD standard deviation, MMSE mini-mental state examination, SNSB Seoul neuropsychological screening battery, DTI diffusion-tensor image, FA functional anisotropy, MD mean diffusivity, Ref reference.