Table 2.

The results of modelling cognitive status (CNC vs AD) and physical activity using Model 1–4 with an adjustment for age, sex, and education.

	Dependent variable: cognitive status (CNC vs AD)
	Model 1	Model 2	Model 3	Model 4
Intercept	7.608**	6.549*	10.588**	12.368***
	(3.567)	(3.615)	(4.139)	(4.591)
Age	−0.051	−0.040	− 0.072*	− 0.089*
	(0.038)	(0.039)	(0.043)	(0.047)
Sex	2.134***	2.111***	2.527***	2.637***
	(0.554)	(0.553)	(0.624)	(0.676)
Education	− 0.224**	− 0.213**	− 0.167*	− 0.174*
	(0.091)	(0.091)	(0.092)	(0.095)
${\overline{X}}_i$	− 0.005***	NA	NA	NA
	(0.002)
$X_i(t)$	NA	$\widehat{\beta }{(t)}^{\ast \ast }$	NA	NA
$Q_i(p)$	NA	NA	$\widehat{\beta }{(p)}^{\ast \ast }$	NA
$Q_i(t,p)$	NA	NA	NA	$\widehat{\beta }{(t,p)}^{\ast \ast \ast }$
Observations	92	92	92	92
cvAUC	0.781	0.773	0.792	0.811

The standard deviation of the estimated coefficients for the scalar predictors are indicated in the parenthesis. Predictors: model 1-scalar average PA, model 2–diurnal PA curves, model 3-quantile functions, model 4-SOTDR with time-by-distribution data objects.

*$p<0.1$; **$p<0.05$; ***$p<0.01$.