Table 5 Prediction of morning plasma AD biomarker levels following overnight sleep using the null model and the neuro-glymphatic model with PhysioS*

From: The glymphatic system clears amyloid beta and tau from brain to plasma in humans

Model

Predictor Group

Predictor

Amyloid Positive

Amyloid Negative

Estimate

DF

CI

P

Estimate

DF

CI

P

Null Model

Fixed Effects

Age

-0.827

42

−1.856 – 0.203

0.113

0.075

108

-0.074 – 0.223

0.323

Sex (male)

4.504

42

0.026 – 8.982

0.049

0.652

108

-0.587 – 1.890

0.299

APOE-ε4 (positive)

-2.148

42

-6.949 – 2.652

0.372

−1.077

108

-2.365 – 0.212

0.101

Circadian

0.039

42

-0.055 – 0.133

0.411

0.003

108

-0.008 – 0.014

0.590

Study (Villages)

-

-

-

-

−1.309

108

-2.986 – 0.367

0.124

Evening Biomarker Effects

ConcPM

0.929

42

0.882 – 0.976

<0.001

0.959

108

0.933 – 0.986

<0.001

Random Effects

σ2

37.327

   

3.842

   

τ11 pid Ab40

1.398

   

0.000

   

τ11 pid Ab42

1.289

   

18.558

   

τ11 Aggregate

τ11 pid np-tau181

34.643

   

193.466

   

τ11 pid np-tau217

   

352.682

   

τ11 pid p-tau181

   

1.482

   

Neuro-Glymphatic Model

Fixed Effects

Age

2.992

36

1.552 – 4.431

<0.001

0.137

98

0.050 – 0.224

0.002

Sex (male)

-5.951

36

−10.868 – −1.034

0.019

0.768

98

0.057 – 1.479

0.034

APOE-ε4 (positive)

-6.993

36

-9.946 – -4.039

<0.001

-0.580

98

−1.314 – 0.154

0.120

Circadian

-0.158

36

-0.236 – -0.079

<0.001

0.008

98

0.002 – 0.015

0.010

Study (Villages)

-

-

-

-

0.274

98

-0.831 – 1.379

0.623

Evening Biomarker Effects

ConcPM

4.067

36

3.129 – 5.004

<0.001

1.043

98

0.919 – 1.167

<0.001

Predictor Effects

Aβ40: PhysioS

−150.535

36

-290.48 – −10.586

0.036

5.190

98

-2.328 – 12.71

0.174

Aβ42: PhysioS

−152.504

36

-290.41 – −14.597

0.031

20.988

98

9.22 – 32.758

0.001

np-tau181: PhysioS

−173.692

36

-312.086 – -35.298

0.015

31.638

98

23.24 – 40.04

<0.001

np-tau217: PhysioS

36

44.178

98

31.69 – 56.67

<0.001

p-tau181: PhysioS

36

1.760

98

-0.935 – 4.456

0.198

Evening Biomarker: Predictor Interactions

(Aβ40: PhysioS) ConcPM

-4.284

36

-5.550 – -3.018

<0.001

-0.258

98

-0.46 – -0.055

0.013

(Aβ42: PhysioS) ConcPM

-4.371

36

-5.678 – -3.064

<0.001

-0.631

98

-0.91 – -0.353

<0.001

(np-tau181:PhysioS) ConcPM

-3.742

36

-4.862 – -2.621

<0.001

-0.559

98

-0.72 – -0.399

<0.001

np-tau217 PhysioS) ConcPM

-0.605

98

-0.78 – -0.430

<0.001

(p-tau181: PhysioS) ConcPM

-0.528

98

-0.71 – -0.351

<0.001

Random Effects

σ2

17.538

   

0.608

   

τ11 pid Ab40

12.406

   

6.374

   

τ11 pid Ab42

12.432

   

19.957

   

τ11 Aggregate

τ11 pid np-tau181

2.105

   

71.422

   

τ11 pid np-tau217

   

157.855

   

τ11 pid p-tau181

   

0.345

   
  1. *DF, degrees of freedom; CI, 95% confidence interval; P, p-value; PhysioS, single linear predictor constructed from parenchymal resistance RP, cardio cerebrovascular metrics and EEG powerbands; ConcPM, evening plasma level; UF, University of Florida study site; EEG, Electroencephalography; σ2, residual variance; τ11, random slope variance; Two-sided p-values were calculated using a t-distribution with Satterthwaite degrees of freedom and were not adjusted for multiple comparisons. Bolded P values reflect significant associations at P < 0.05.