Table 4 Full model predictors and optimal weights predicting morning plasma AD biomarker levels in sleep and sleep deprivation conditions*

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

Condition and Predictor

Amyloid Status

Predictors

Estimate

CI

P

Sleep Condition

PhysioS

Positive

Rp

0.968

0.177–0.998

0.036

PTT (NREM)

−0.034

−0.396–0.132

0.388

HRV (NREM)

−0.068

−0.464–−0.014

0.030

EEG Delta Power (NREM)

0.086

−0.198–0.705

0.399

EEG Theta Power (NREM)

0.162

−0.261–0.604

0.195

EEG Theta Power (REM)

−0.146

−0.547–0.496

0.436

EEG Beta Power (REM)

−0.041

−0.671–0.095

0.264

Negative

Rp

0.963

0.948–0.977

<0.001

PTT (NREM)

−0.058

−0.200–−0.028

0.006

HRV (NREM)

0.114

−0.003–0.178

0.060

EEG Delta Power (NREM)

−0.041

−0.149–−0.010

0.026

EEG Theta Power (NREM)

0.169

0.083–0.218

<0.001

EEG Theta Power (REM)

0.087

0.001–0.133

0.058

EEG Beta Power (REM)

−0.135

−0.250–−0.093

<0.001

Sleep Condition

HypnoS

Positive

NREM Duration

−0.182

−0.629–0.108

0.166

REM Duration

−0.519

−0.829–0.337

0.209

N1 Duration

0.834

0.366–0.990

0.010

WASO Duration

0.044

−0.396–0.477

0.620

Negative

NREM Duration

−0.090

−0.239–−0.054

0.001

REM Duration

0.525

0.423–0.585

<0.001

N1 Duration

0.801

0.754–0.855

<0.001

WASO Duration

0.273

0.131–0.344

0.019

Sleep Deprivation/Wake

Physiow

Positive

Rp

0.621

−0.365–0.891

0.087

PTT

0.190

−0.134–0.505

0.390

HRV

−0.242

−0.450–0.106

0.166

EEG Delta Power

−0.487

−0.831–−0.124

0.027

EEG Theta Power

−0.382

−0.670–0.584

0.339

EEG Beta Power

0.371

−0.036–0.690

0.071

Negative

Rp

0.626

0.519–0.651

<0.001

PTT

−0.004

−0.127–0.029

0.461

HRV

0.136

0.058–0.173

0.028

EEG Delta Power

−0.658

−0.717–−0.634

<0.001

EEG Theta Power

−0.391

−0.473–−0.359

<0.001

EEG Beta Power

0.062

−0.026–0.095

0.170

  1. *CI, 95% Confidence Interval; Rp, Parenchymal Resistance; PTT, Pulse Transit Time; HRV, Heart Rate Variability; EEG, Electroencephalography; PhysioS and PhysioW, single predictors constructed from the linear combination of the full model predictors during the sleep and sleep deprivation (wake) conditions, respectively, for both amyloid positive and negative participants separately; HypnoS, single predictor constructed from the full model hypnogram features during the sleep condition for both amyloid positive and negative participants separately. The estimated weights were selected using optimization to maximize the log likelihood of the production/clearance model. Confidence intervals were calculated using the 95% quantiles of 1000 nonparametric bootstrap replicates. Two-sided p-values were obtained from the nonparametric bootstrap by calculating the proportion of bootstrap estimates exceeding the observed effect and multiplying by two. P-values were not adjusted for multiple comparisons. Bolded P values reflect significant associations at P < 0.05.
Back to article page