Table 3.

Linear model of predictors of change in SMN connectivity (R² = 0.166; F(5,62) = 3.065; p = 0.016) and DAN–aDMN connectivity (R² = 0.088; F(5,62) = 1.023; p = 0.412).

Model terms	B [95% CI]	SE B (HC4)	t	p
SMN change
Constant	− 0.071 [− 0.643, 0.500]	0.286	− 0.250	0.803
Program	0.053 [− 0.021, 0.127]	0.037	1.430	0.158
Cognitive reserve (centered)	− 0.016 [− 0.035, 0.003]	0.010	− 1.709	0.093
Moderator CR*program	0.045 [0.017, 0.074]	0.014	3.169	0.002
Age	0.001 [− 0.007, 0.009]	0.004	0.285	0.777
Sex	− 0.016 [− 0.100, 0.069]	0.042	− 0.375	0.709
DAN–aDMN change
Constant	− 0.064 [− 0.353, 0.225]	0.145	− 0.441	0.661
Program	0.062 [0.005, 0.119]	0.029	2.160	0.035
Cognitive reserve (centered)	− 0.001 [− 0.018, 0.015]	0.008	− 0.182	0.856
Moderator CR*program	0.008 [− 0.014, 0.029]	0.011	0.717	0.476
Age	0.001 [− 0.003, 0.005]	0.002	0.315	0.754
Sex	− 0.021 [− 0.102, 0.059]	0.040	− 0.523	0.603

Results are unstandardized beta coefficients with 95% confidence intervals from moderation models estimating the association of program, CR (years of education), and their interaction with rs-FC change. Note: only continuous variables that contributed to the outcomes were centered. Cribari-Neto model was used for standard error of variance and F-statistics. Variable program coded as LAU: 0, DI: 1; SMN inter-network rs-FC of the sensorimotor network; DAN–aDMN between-network rs-FC of dorsal attention network and anterior part of the default mode network.