Article Figures & Data
Figures
- Figure 1.
Participant flow and at which time point data were collected.
- Figure 2.
Wave by individual peak α frequency interaction plot. Plot depicting the two-way interaction wave × individual peak α frequency reported in Table 6 with 95% confidence interval error bars. Time 1 cognition assessed at MIDUS2 Cognitive Project, and time 2 cognition was assessed at the MIDUS 3 Cognitive Project.
- Figure 3.
Wave by aperiodic exponent by individual peak α frequency interaction plot. Plot depicting the three-way interaction wave × aperiodic exponent × individual peak α frequency reported in Table 9, with wave depicted as the estimated change in cognitive function between the M2 and M3 Cognitive Projects.
Tables
Age in years: MIDUS 2 Neuroscience Project Sex M(SD) = 55.10 (10.71) Male 94 (40.0%) Female 141 (60.0%) 36–49 85 (36.2%) Race 50–65 105 (44.7%) White 173 (73.6%) 66–83 45 (19.1%) Total black, indigenous, and people of color (BIPOC) 62 (26.4%) Education High school or less 67 (29.3%) Some college 70 (30.6%) Bachelor's or higher 92 (40.2%) Hypotheses Analytic plana Results H1: Cognitive function at time 1 (indexed by the BTACT composite score) will be negatively associated with age.H2: Within-person changes in cognitive function (indexed by time 1 to time 2 changes in BTACT composite score) will be moderated by age, such that older age will be associated with greater decline in cognitive function between time 1 and 2. Multilevel linear model controlling for education, sex, and lag between waves in years as follows:
Equation 1
Level 1:
Level 2:
H1: Negative coefficient for the main effect of age at time 1 (i.e.,
H2: Negative coefficient for the interaction between
Age and wave (i.e.,
H1: Supported for BTACT composite and both executive functioning and episodic memory subfactors.H2: Not supported. H3: Cognitive function at time 1 (indexed by the BTACT composite score) will be positively associated with aperiodic exponent, such that flatter spectra will be associated with poorer cognitive function.H4: Within-person changes in cognitive function (indexed by time 1 to time 2 changes in BTACT composite score) will be moderated by the aperiodic exponent, such that flatter spectra at time 1 will be associated with greater decline in cognitive function between time 1 and 2.H5: Cognitive function at time 1 (indexed by the BTACT composite score) will be positively associated with IAPF, such that lower IAPF will be associated with poorer cognitive function.H6: Within-person changes in cognitive function (indexed by time 1 to time 2 changes in BTACT composite score) will be moderated by IAPF, such that lower IAPF at time 1 will be associated with greater decline in cognitive function between time 1 and 2. Multilevel linear model controlling for education, sex, and lag between waves in years. The placeholder “EEG” is used to represent the different EEG metrics included separately in each of the models as per hypotheses.
Equation 2
Level 1:
Level 2:
H3: Significant positive coefficient for the EEG metric (i.e.,
H4: Significant coefficient for the EEG by Wave interaction (i.e.,
H5: Significant positive coefficient for the EEG metric (i.e.,
H6: Significant coefficient for the EEG by wave interaction (i.e.,
H3: Supported for BTACT composite and executive functioning subfactor.H4: Not supported.H5: Not supported.H6: Supported for BTACT composite. H7: The relationships between the aperiodic exponent, individual α peak frequency, and cognitive function outlined in H3–H6 will be moderated by age, such that greater cognitive decline will be observed in older-aged participants with flatter aperiodic exponents and lower IAPF. This hypothesis can be broken down into 4 parts (a–d) as described in the “Analytic Plan”. Multilevel linear model controlling for education, sex, and lag between waves in years. The placeholder “EEG” is used to represent the different EEG metrics we aim to include in each of the models.
Equation 3
Level 1:
Level 2:
H7a: Significant coefficient for the interaction between the EEG by Age (i.e.,
H7b: Significant coefficient for the interaction between the EEG by Age (i.e.,
H7c: Significant coefficient for the interaction between the EEG by Age by Wave (i.e.,
H7d: Significant coefficient for the interaction between the EEG by Age by Wave (i.e.,
H7a: Not supported. H7b: Not supported. H7c: Not supported. H7d: Not supported. ↵a The exact preregistered analyses were overly complicated and conservative by including lag as two separate variables (instead of a linear addition into a single variable), as well as the interaction between covariates and wave. Additionally, after extensive testing for possible interactions with education and finding none, we decided to include education as a more appropriate covariate than race. We report the preregistered analyses, which are consistent with these findings, on OSF at https://doi.org/10.17605/OSF.IO/SR4MB.
Mean (SD) n 2. Total lag 3. M2 Episodic memory 4. M3 Episodic memory 5. M2 Executive function 6. M3 Executive function 7. M2 BTACT composite 8. M3 BTACT composite 9. Aperiodic exponent 10. IAPF 1. Age (MIDUS 2 Neuroscience) 55.10 (10.71) n = 235 −0.12
[−0.25, 0.01]
p = 0.076
n = 235
−0.18
[−0.30, −0.05]
p = 0.010
n = 228
−0.24
[−0.36, −0.11]
p = 0.001
n = 223
−0.35
[−0.46, −0.23]
p < 0.001
n = 228
−0.32
[−0.43, −0.20]
p < 0.001
n = 230
−0.36
[−0.47, −0.24]
p < 0.001
n = 224
−0.34
[−0.45, −0.21]
p < 0.001
n = 219
−0.29
[−0.40, −0.17]
p < 0.001
n = 235
−0.23
[−0.34, −0.10]
p = 0.001
n = 235
2. Total Lag (M2 to M3 Cognitive Projects, years) 9.71 (0.92) n = 235 - −0.06
[−0.19, 0.07]
p = 0.389
n = 228
−0.07
[−0.20, 0.06]
p = 0.362
n = 223
−0.17
[−0.29, −0.04]
p = 0.018
n = 228
−0.30
[−0.41, −0.17]
p < 0.001
n = 230
−0.16
[−0.29, −0.03]
p = 0.021
n = 224
−0.29
[−0.41, −0.17]
p < 0.001
n = 219
0.14
[0.01, 0.26]
p = 0.049
n = 235
−0.06
[−0.18, 0.07]
p = 0.416
n = 235
3. Episodic memory MIDUS 2 0.00 (0.93) n = 228 - 0.56
[0.46, 0.64]
p < 0.001
n = 217
0.36
[0.24, 0.47]
p < 0.001
n = 224
0.32
[0.19, 0.43]
p < 0.001
n = 223
0.56
[0.47, 0.65]
p < 0.001
n = 224
0.41
[0.29, 0.52]
p < 0.001
n = 213
0.03
[−0.10, 0.16]
p = 0.639
n = 228
0.16
[0.03, 0.28]
p = 0.023
n = 228
4. Episodic memory MIDUS 3 −0.11 (1.10) n = 223 - 0.26
[0.13, 0.38]
p < 0.001
n = 217
0.37
[0.25, 0.48]
p < 0.001
n = 219
0.39
[0.27, 0.50]
p < 0.001
n = 213
0.59
[0.49, 0.67]
p < 0.001
n = 219
0.01
[−0.12, 0.14]
p = 0.867
n = 223
0.19
[0.06, 0.31]
p = 0.007
n = 223
5. Executive function MIDUS 2 0.01 (0.65) n = 228 - 0.69
[0.62, 0.76]
p < 0.001
n = 223
0.97
[0.96, 0.98]
p < 0.001
n = 224
0.68
[0.60, 0.74]
p < 0.001
n = 213
0.13
[0.00, 0.26]
p = 0.054
n = 228
0.14
[0.01, 0.27]
p = 0.045
n = 228
6. Executive function MIDUS 3 −0.44 (0.74) n = 230 - 0.69
[0.61, 0.75]
p < 0.001
n = 219
0.97
[0.96, 0.98]
p < 0.001
n = 219
0.04
[−0.09, 0.17]
p = 0.566
n = 230
0.19
[0.06, 0.31]
p = 0.006
n = 230
7. BTACT composite MIDUS 2 0.01 (0.62) n = 224 - 0.70
[0.63, 0.77]
p < 0.001
n = 209
0.13
[0.00, 0.26]
p = 0.062
n = 224
0.16
[0.03, 0.29]
p = 0.020
n = 224
8. BTACT composite MIDUS 3 −0.37 (0.71) n = 219 - 0.04
[−0.09, 0.17]
p = 0.566
n = 219
0.21
[0.08, 0.34]
p = 0.003
n = 219
9. Aperiodic exponent 1.20 (0.28) n = 235 - −0.16
[−0.28, −0.03]
p = 0.020
n = 235
10. IAPF 9.60 (0.98) n = 229 - Bold values are significant at p < .05.
Predictors Episodic memory Executive functioning BTACT composite Estimates 95% CI P Estimates 95% CI p Estimates 95% CI p Intercept (M2) −0.43 −0.60 to −0.26 <0.001 −0.07 −0.19 to 0.05 0.256 −0.12 −0.23 to −0.01 0.028 MIDUS wave −0.13 −0.26 to −0.00 0.043 −0.43 −0.50 to −0.36 <0.001 −0.38 −0.45 to −0.31 <0.001 Age −0.02 −0.03 to −0.01 0.001 −0.02 −0.03 to −0.02 <0.001 −0.02 −0.03 to −0.02 <0.001 Sex 0.74 0.53 to 0.94 <0.001 0.11 −0.03 to 0.26 0.128 0.22 0.08 to 0.35 0.001 Education 0.17 0.04 to 0.29 0.008 0.13 0.05 to 0.21 0.001 0.14 0.06 to 0.21 0.001 Lag between waves −0.10 −0.21 to 0.01 0.073 −0.21 −0.29 to −0.14 <0.001 −0.20 −0.27 to −0.13 <0.001 MIDUS wave × age −0.01 −0.02 to 0.00 0.096 0.00 −0.01 to 0.01 0.994 −0.00 −0.01 to 0.01 0.707 Random effects σ2 0.46 0.16 0.14 τ00 0.38 M2ID 0.22 M2ID 0.19 M2ID ICC 0.45 0.59 0.58 N 234 M2ID 235 M2ID 234 M2ID Observations 451 458 443 Marginal R2/conditional R2 0.195/0.559 0.292/0.707 0.311/0.708 Repeated measured cognitive data nested within participant, indicated by M2ID. Sex coded as 0 = male, 1 = female. Education coded as −1 = high school or less, 0 = some college, 1 = bachelor's degree or higher. MIDUS wave coded as 0 = MIDUS 2, 1 = MIDUS 3. Years between waves (i.e., between MIDUS 2 and MIDUS 3 Cognitive Projects) and age at M2 Neuroscience Project are mean centered. Empirical Bayes slope estimation used (Bates et al., 2015).
Bold values are significant at p < .05.
Predictors Episodic memory Executive functioning BTACT composite Estimates 95% CI p Estimates 95% CI p Estimates 95% CI p Intercept (M2) −0.41 −0.59 to −0.23 <0.001 −0.05 −0.18 to 0.08 0.450 −0.11 −0.23 to 0.01 0.085 MIDUS wave −0.13 −0.26 to 0.00 0.051 −0.43 −0.51 to −0.36 <0.001 −0.38 −0.45 to −0.31 <0.001 Exponent 0.26 −0.19 to 0.71 0.254 0.40 0.09 to 0.71 0.012 0.39 0.10 to 0.68 0.009 Sex 0.71 0.49 to 0.93 <0.001 0.08 −0.08 to 0.24 0.306 0.19 0.04 to 0.34 0.012 Education 0.17 0.04 to 0.30 0.008 0.13 0.04 to 0.21 0.003 0.13 0.05 to 0.21 0.002 Lag between waves −0.08 −0.19 to 0.04 0.187 −0.19 −0.28 to −0.11 <0.001 −0.18 −0.26 to −0.10 <0.001 Wave × exponent −0.02 −0.48 to 0.44 0.926 −0.21 −0.47 to 0.05 0.120 −0.16 −0.42 to 0.09 0.204 Random effects σ2 0.47 0.15 0.14 τ00 0.44 M2ID 0.28 M2ID 0.25 M2ID ICC 0.48 0.65 0.65 N 234 M2ID 235 M2ID 234 M2ID Observations 451 458 443 Marginal R2/Conditional R2 0.136/0.551 0.179/0.710 0.182/0.710 Repeated measured cognitive data nested within participant, indicated by M2ID. Sex coded as 0 = male, 1 = female. Education coded as −1 = high school or less, 0 = some college, 1 = bachelor's degree or higher. MIDUS wave coded as 0 = MIDUS 2, 1 = MIDUS 3. Years between waves (i.e., between MIDUS 2 and MIDUS 3 Cognitive Projects) and age at M2 Neuroscience Project are mean centered. Empirical Bayes slope estimation used (Bates et al., 2015).
Bold values are significant at p < .05.
Predictors Episodic memory Executive functioning BTACT composite Estimates 95% CI P Estimates 95% CI p Estimates 95% CI p Intercept (M2) −0.40 −0.57 to −0.22 <0.001 −0.03 −0.16 to 0.09 0.596 −0.09 −0.21 to 0.03 0.146 MIDUS wave −0.13 −0.25 to 0.00 0.052 −0.43 −0.51 to −0.36 <0.001 −0.38 −0.45 to −0.31 <0.001 IAPF 0.12 −0.00 to 0.25 0.055 0.07 −0.02 to 0.16 0.115 0.08 −0.00 to 0.16 0.058 Sex 0.68 0.47 to 0.89 <0.001 0.06 −0.10 to 0.21 0.469 0.16 0.02 to 0.31 0.030 Education 0.17 0.05 to 0.30 0.007 0.14 0.05 to 0.22 0.002 0.13 0.05 to 0.21 0.001 Lag between waves −0.06 −0.17 to 0.06 0.326 −0.17 −0.26 to −0.09 <0.001 −0.16 −0.24 to −0.08 <0.001 Wave × IAPF 0.07 −0.06 to 0.20 0.281 0.07 −0.00 to 0.15 0.061 0.07 0.00 to 0.14 0.047 Random effects σ2 0.47 0.15 0.13 τ00 0.42 M2ID 0.28 M2ID 0.25 M2ID ICC 0.47 0.64 0.65 N 234 M2ID 235 M2ID 234 M2ID Observations 451 458 443 Marginal R2/conditional R2 0.156/0.553 0.188/0.711 0.196/0.715 Repeated measured cognitive data nested within participant, indicated by M2ID. Sex coded as 0 = male, 1 = female. Education coded as −1 = high school or less, 0 = some college, 1 = bachelor's degree or higher. MIDUS wave coded as 0 = MIDUS 2, 1 = MIDUS 3. Years between waves (i.e., between MIDUS 2 and MIDUS 3 Cognitive Projects) and age at M2 Neuroscience Project are mean centered. Empirical Bayes slope estimation used (Bates et al., 2015).
Bold values are significant at p < .05.
Predictors Episodic memory Executive functioning BTACT composite Estimates 95% CI P Estimates 95% CI p Estimates 95% CI p Intercept (M2) −0.44 −0.62 to −0.26 <0.001 −0.06 −0.18 to 0.06 0.340 −0.12 −0.23 to −0.00 0.046 MIDUS wave −0.11 −0.24 to 0.03 0.122 −0.43 −0.50 to −0.35 <0.001 −0.38 −0.46 to −0.31 <0.001 Exponent 0.07 −0.38 to 0.52 0.748 0.16 −0.14 to 0.46 0.304 0.15 −0.13 to 0.43 0.290 Age −0.02 −0.03 to −0.01 0.002 −0.02 −0.03 to −0.01 <0.001 −0.02 −0.03 to −0.01 <0.001 Sex 0.73 0.52 to 0.94 <0.001 0.11 −0.04 to 0.26 0.140 0.22 0.08 to 0.36 0.002 Education 0.16 0.04 to 0.29 0.010 0.13 0.05 to 0.21 0.002 0.13 0.06 to 0.21 0.001 Lag between waves −0.10 −0.21 to 0.01 0.080 −0.22 −0.29 to −0.14 <0.001 −0.20 −0.27 to −0.13 <0.001 Wave × exponent −0.14 −0.62 to 0.33 0.557 −0.23 −0.50 to 0.05 0.104 −0.19 −0.45 to 0.08 0.166 Wave × age −0.01 −0.02 to 0.00 0.122 −0.00 −0.01 to 0.01 0.687 −0.00 −0.01 to 0.00 0.476 Exponent × age −0.01 −0.06 to 0.04 0.677 0.01 −0.02 to 0.04 0.632 0.01 −0.02 to 0.03 0.695 Wave × exponent × age 0.03 −0.02 to 0.08 0.201 0.00 −0.02 to 0.03 0.755 0.00 −0.02 to 0.03 0.858 Random effects σ2 0.46 0.16 0.14 τ00 0.39 M2ID 0.22 M2ID 0.19 M2ID ICC 0.46 0.59 0.58 N 234 M2ID 235 M2ID 234 M2ID Observations 451 458 443 Marginal R2/conditional R2 0.195/0.563 0.294/0.710 0.311/0.710 Repeated measured cognitive data nested within participant, indicated by M2ID. Sex coded as 0 = male, 1 = female. Education coded as −1 = high school or less, 0 = some college, 1 = bachelor's degree or higher. MIDUS wave coded as 0 = MIDUS 2, 1 = MIDUS 3. Years between waves (i.e., between MIDUS 2 and MIDUS 3 Cognitive Projects) and age at M2 Neuroscience Project are mean centered. Empirical Bayes slope estimation used (Bates et al., 2015).
Bold values are significant at p < .05.
Predictors Episodic memory Executive functioning BTACT Estimates 95% CI p Estimates 95% CI p Estimates 95% CI p Intercept (M2) −0.42 −0.59 to −0.24 <0.001 −0.06 −0.18 to 0.06 0.360 −0.11 −0.22 to 0.00 0.052 MIDUS wave −0.13 −0.26 to 0.00 0.057 −0.44 −0.52 to −0.36 <0.001 −0.39 −0.46 to −0.32 <0.001 IAPF 0.08 −0.05 to 0.21 0.225 0.01 −0.07 to 0.10 0.797 0.02 −0.06 to 0.10 0.610 Age −0.02 −0.03 to −0.01 0.005 −0.02 −0.03 to −0.02 <0.001 −0.02 −0.03 to −0.02 <0.001 Sex 0.72 0.51 to 0.93 <0.001 0.11 −0.04 to 0.25 0.145 0.21 0.08 to 0.35 0.002 Education 0.17 0.04 to 0.29 0.009 0.14 0.06 to 0.22 0.001 0.14 0.06 to 0.21 <0.001 Lag between waves −0.09 −0.20 to 0.02 0.116 −0.21 −0.29 to −0.13 <0.001 −0.19 −0.26 to −0.12 <0.001 Wave × IAPF 0.05 −0.09 to 0.18 0.483 0.07 −0.01 to 0.15 0.071 0.07 −0.01 to 0.14 0.076 Wave × age −0.01 −0.02 to 0.00 0.165 0.00 −0.01 to 0.01 0.741 −0.00 −0.01 to 0.01 0.994 IAPF × age 0.00 −0.01 to 0.01 0.836 0.00 −0.00 to 0.01 0.350 0.00 −0.00 to 0.01 0.374 Wave × IAPF × age 0.00 −0.01 to 0.01 0.896 −0.00 −0.01 to 0.00 0.301 −0.00 −0.01 to 0.00 0.379 Random effects σ2 0.47 0.15 0.14 τ00 0.38 M2ID 0.22 M2ID 0.19 M2ID ICC 0.45 0.59 0.58 N 234 M2ID 235 M2ID 234 M2ID Observations 451 458 443 Marginal R2/conditional R2 0.203/0.559 0.298/0.713 0.318/0.715 Repeated measured cognitive data nested within participant, indicated by M2ID. Sex coded as 0 = male, 1 = female. Education coded as −1 = high school or less, 0 = some college, 1 = bachelor's degree or higher. MIDUS wave coded as 0 = MIDUS 2, 1 = MIDUS 3. Years between waves (i.e., between MIDUS 2 and MIDUS 3 Cognitive Projects) and age at M2 Neuroscience Project are mean centered. Empirical Bayes slope estimation used (Bates et al., 2015).
Bold values are significant at p < .05.
- Table 9.
Multilevel models examine the interaction between aperiodic exponent and individual peak α frequency
Predictors Episodic memory Executive functioning BTACT Estimates 95% CI p Estimates 95% CI p Estimates 95% CI p Intercept (M2) −0.43 −0.60 to −0.25 <0.001 −0.07 −0.19 to 0.05 0.228 −0.13 −0.24 to −0.02 0.024 MIDUS wave −0.12 −0.25 to 0.01 0.069 −0.41 −0.49 to −0.34 <0.001 −0.37 −0.44 to −0.30 <0.001 Exponent 0.06 −0.41 to 0.53 0.802 0.13 −0.19 to 0.45 0.417 0.13 −0.17 to 0.42 0.393 IAPF 0.07 −0.06 to 0.20 0.269 0.02 −0.07 to 0.11 0.634 0.03 −0.05 to 0.11 0.482 Age −0.02 −0.03 to −0.01 <0.001 −0.02 −0.03 to −0.02 <0.001 −0.02 −0.03 to −0.02 <0.001 Sex 0.73 0.52 to 0.94 <0.001 0.11 −0.03 to 0.26 0.135 0.22 0.08 to 0.35 0.002 Education 0.17 0.05 to 0.30 0.007 0.14 0.06 to 0.22 0.001 0.14 0.07 to 0.22 <0.001 Lag between waves −0.09 −0.21 to 0.02 0.116 −0.21 −0.29 to −0.13 <0.001 −0.20 −0.27 to −0.12 <0.001 Wave × exponent 0.09 −0.39 to 0.58 0.714 −0.08 −0.36 to 0.19 0.546 −0.03 −0.29 to 0.23 0.826 Wave × IAPF 0.07 −0.06 to 0.20 0.314 0.06 −0.02 to 0.13 0.146 0.06 −0.01 to 0.13 0.118 Exponent × IAPF −0.07 −0.51 to 0.38 0.773 −0.15 −0.44 to 0.15 0.327 −0.14 −0.42 to 0.13 0.316 Wave × exponent × IAPF 0.20 −0.26 to 0.67 0.397 0.34 0.07 to 0.61 0.013 0.33 0.08 to 0.59 0.010 Random effects σ2 0.47 0.15 0.13 τ00 0.38 M2ID 0.23 M2ID 0.19 M2ID ICC 0.44 0.60 0.59 N 234 M2ID 235 M2ID 234 M2ID Observations 451 458 443 Marginal R2/conditional R2 0.202/0.557 0.302/0.722 0.323/0.724 Repeated measured cognitive data nested within participant, indicated by M2ID. Sex coded as 0 = male, 1 = female. Education coded as −1 = high school or less, 0 = some college, 1 = bachelor's degree or higher. MIDUS wave coded as 0 = MIDUS 2, 1 = MIDUS 3. Years between waves (i.e., between MIDUS 2 and MIDUS 3 Cognitive Projects) and age at M2 Neuroscience Project are mean centered. Empirical Bayes slope estimation used (Bates et al., 2015).
Bold values are significant at p < .05.
- Table 10.
Examining the wave × aperiodic exponent × individual peak α frequency interaction through the slope of the change in cognitive function over waves for each BTACT measure at high and low levels of each EEG metric
Change in BTACT episodic memory factor from M2 to M3 Slope of IAPF 95% CI p-value Low exponent (−1 SD) 0.01 −0.18 to 0.20 0.903 High exponent (+1 SD) 0.12 −0.05 to 0.30 0.173 Slope of exponent 95% CI p-value Low IAPF (−1 SD) −0.11 −0.67 to 0.45 0.709 High IAPF (+1 SD) 0.29 −0.47 to 1.05 0.457 Change in BTACT executive functioning factor from M2 to M3 Slope of IAPF 95% CI p-value Low exponent (−1 SD) −0.04 −0.15 to 0.07 0.482 High exponent (+1 SD) 0.15 0.05 to 0.25 0.004 Slope of exponent 95% CI p-value Low IAPF (−1 SD) −0.42 −0.75 to −0.09 0.013 High IAPF (+1 SD) 0.25 −0.17 to 0.67 0.244 Change in BTACT overall composite from M2 to M3 Slope of IAPF 95% CI p-value Low exponent (−1 SD) −0.04 −0.14 to 0.07 0.502 High exponent (+1 SD) 0.15 0.05 to 0.25 0.002 Slope of exponent 95% CI p-value Low IAPF (−1 SD) −0.36 −0.67 to −0.04 0.025 High IAPF (+1 SD) 0.30 −0.11 to 0.70 0.149 Bold values are significant at p < .05.
