Unveiling the sound of the cognitive status: Machine Learning-based speech analysis in the Alzheimer’s disease spectrum

Table 8 Regression model results for the study sample, controlling for sex, age, and years of formal education

Composite	Condition	CDR	Coef.^a	p-value	_95%CI
Attention	SCD	0.0	1.19	< 0.001	[1.10, 1.29]
	MCI	0.5	− 0.73	< 0.001	[− 0.83, − 0.62]
	Mild ADD	1.0	− 1.15	< 0.001	[− 1.26, − 1.04]
	Moderate ADD	2.0	− 1.47	< 0.001	[− 1.60, − 1.33]
Executive function	SCD	0.0	1.48	< 0.001	[1.39, 1.57]
	MCI	0.5	− 0.91	< 0.001	[− 1.01, − 0.81]
	Mild ADD	1.0	− 1.58	< 0.001	[− 1.69, − 1.48]
	Moderate ADD	2.0	− 1.90	< 0.001	[− 2.03, − 1.77]
Language	SCD	0.0	1.15	< 0.001	[1.05, 1.24]
	MCI	0.5	− 0.63	< 0.001	[− 0.73, − 0.53]
	Mild ADD	1.0	− 1.31	< 0.001	[− 1.42, − 1.20]
	Moderate ADD	2.0	− 1.67	< 0.001	[− 1.81, − 1.54]
Memory	SCD	0.0	1.53	< 0.001	[1.43, 1.62]
	MCI	0.5	− 1.01	< 0.001	[− 1.12, − 0.91]
	Mild ADD	1.0	− 1.78	< 0.001	[− 1.89, − 1.67]
	Moderate ADD	2.0	− 1.95	< 0.001	[− 2.08, − 1.82]
Visuospatial function	SCD	0.0	1.22	< 0.001	[1.12, 1.33]
	MCI	0.5	− 0.73	< 0.001	[− 0.84, − 0.62]
	Mild ADD	1.0	− 1.49	< 0.001	[− 1.61, − 1.37]
	Moderate ADD	2.0	− 1.88	< 0.001	[− 2.02, − 1.73]

Abbreviations: CDR, Clinical Dementia Rating; CI, confidence interval
^aCoefficients of the regression models considering the composite value as the dependent variable and the dummy coding of the clinical diagnosis as the independent variable. The models were adjusted for sex, age, and educational level. The coefficients indicate the average value of the composite associated with each group adjusting for the covariates

ISSN: 1758-9193