Skip to main content
. 2019 May 31;11:126. doi: 10.3389/fnagi.2019.00126

Table 2.

Multilevel anlaysis of the training data (n-back task).

MODEL 1
MODEL 2
MODEL 3
MODEL 4
MODEL 5
MODEL 6
+ change in time in Nback task + age as a predictor + sex as a predictor + education level score as a predictor + occupational activity as a predictor + Initial OSPAN






b (s.e.) b (s.e.) b (s.e.) b (s.e.) b (s.e.) b (s.e.)
Fixed effects
Intercept 2.165 (0.088)*** 2.162 (0.093)***   2.216 (0.151)***   2.238 (0.111)***   2.278 (0.158)*** 1.928 (0.143)***
Time – linear (centered at 1st day) 0.031 (0.005)*** 0.028 (0.005)***   0.036 (0.008)***   0.035 (0.006)***   0.037 (0.008)*** 0.016 (0.007)*
Ageˆ 0.010 (0.018)
      n.s.
Sex∼ –0.008 (0.188)
      n.s.
Education level∼ –0.204 (0.186)
      n.s.
Occupational activity∼ –0.131 (0.191)
      n.s.
Initial OSPAN score ∼ (high/low) 0.038 (0.183)*
Time × age –0.002 (0.001)°
Time × sex –0.008 (0.009)
      n.s.
Time × education level –0.011 (0.009)
      n.s.
Time × occupational activity –0.008 (0.010)
      n.s.
Time × initial OSPAN score 0.026 (0.008)**
Random effects ([co-]variances)
Level 2 (between-person)
   Intercept 0.285 (0.073)*** 0.301 (0.079)*** 0.300 (0.006)*** 0.292 (0.076)*** 0.283 (0.075)*** 0.286 (0.075)***
   Time – linear 0.001 (0.001)*** 0.001 (0.001)*** 0.001 (0.001)*** 0.001 (0.001)*** 0.001 (0.001)*** 0.001 (0.001)***
   Intercept and time 0.006 (0.003)* 0.006 (0.003)* 0.006 (0.003)* 0.006 (0.003)* 0.005 (0.003)° 0.004 (0.002)°
Level 1 (within-person)
   Residual 0.151 (0.009)*** 0.149 (0.009)*** 0.150 (0.009)*** 0.151 (0.009)*** 0.153 (0.009)*** 0.149 (0.009)***
   Autocorrelation 0.339 (0.039)*** 0.336 (0.040)*** 0.338 (0.039)*** 0.338 (0.039)*** 0.336 (0.039)*** 0.328 (0.040)***
Model fit
–2 log likelihood (y2) 1069.32 1063.69 1089.51 1088.24 1077.91 1046.37
Akaike’s Information Criterion (AIC) 1083.32 1073.69 1099.51 1098.24 1087.91 1056.37

Results from multilevel modeling. Unstandardized regression coefficients listed with standard errors in parenthesis, °p = 0.1, *p < 0.05, **p < 0.01, and ***p < 0.001. All p-Values are two-tailed except in the case of variances, where one-tiled p-Values are used (because variances are constrained to be non-negative). The predictors are mean centered. The predictors are dichotomous.