TABLE 2.

Demographic, clinical, genetic and biological characteristics of the training and test sets.

		N	Age (years)	Sex (M/F)	Education (years)	MMSE (raw score)	Aβ1-42 (positive/negative)	APOE-ε4 (carriers/ non carriers)
Training set	CN	335	73.5 ± 5.9	46%/54%	16.3 ± 2.6	29.1 ± 1.2	40%/60%	27%/73%
	MCI	537	72.0 ± 7.2	59%/41%	16.0 ± 2.8	27.7 ± 1.8	66%/34%	51%/49%
	AD	171	73.4 ± 8.2	54%/46%	15.5 ± 2.7	23.4 ± 2.0	95%/5%	73%/27%
	Total	1043	72.7 ± 7.0	54%/46%	16.04 ± 2.7	27.4 ± 2.5	62%/38%	46%/54%
	sMCI	271	72.3 ± 7.1	58%/42%	16.1 ± 2.8	28.0 ± 1.7	56%/44%	42%/58%
	pMCI	205	73.1 ± 6.8	59%/41%	15.8 ± 2.8	27.2 ± 1.8	87%/13%	64%/36%
Test Set	CN	440	54 ± 25*	37%/63%*	8.4 ± 5.7*	29.0 ± 1.2	NA	2%/7%
	MCI	283	72.3 ± 7.6	46%/54*	9.2 ± 5.1*	26.3 ± 2.6*	34%/17%	19%/32%*
	AD	44	77.3 ± 7.4*	34%/66%*	5.8 ± 5.3*	21.7 ± 3.8*	NA	0%/5%
	Total	767	62 ± 21*	40%/60%*	8.6 ± 5.4*	27.2 ± 3.0	12%/6%	8%/16%*
	sMCI	152	71.2 ± 7.5	47%/53%	11.5 ± 4.2*	26.7 ± 2.2*	46%/25%	25%/43%
	pMCI	39	69.8 ± 6.4*	49%/51%	11.7 ± 3.9*	25.7 ± 2.4*	44%/5%	33%/26%

Values from CN, MCI and AD contribute to the totals, MCI subpopulations of pMCIs and sMCIs are reported as well. Values marked with * on the test set are significantly different (p-value of ANOVA for continuous variables and chi-square for discrete variables <0.05) from the corresponding values from training set. Abbreviations: M, male; F, female; N, number.