Time-varying correlations between delta EEG power and heart rate variability in midlife women: the SWAN Sleep Study

Abstract

No studies have evaluated the dynamic, time-varying relationship between delta electroencephalographic (EEG) sleep and high frequency heart rate variability (HF-HRV) in women. Delta EEG and HF-HRV were measured during sleep in 197 midlife women (M(age) = 52.1, SD = 2.2). Delta EEG-HF-HRV correlations in nonrapid eye movement (NREM) sleep were modeled as whole-night averages and as continuous functions of time. The whole-night delta EEG-HF-HRV correlation was positive. The strongest correlations were observed during the first NREM sleep period preceding and following peak delta power. Time-varying correlations between delta EEG-HF-HRV were stronger in participants with sleep-disordered breathing and self-reported insomnia compared to healthy controls. The dynamic interplay between sleep and autonomic activity can be modeled across the night to examine within- and between-participant differences including individuals with and without sleep disorders.

Keywords: Delta EEG; HF-HRV; Sleep; Time-varying correlation.

Figure 1

Delta EEG power, HF-HRV, and time-varying correlation for NREM-1, NREM-2 and NREM-3 in the full sample (n=197). Data appear in relative time as opposed to absolute (clock) time in order to account for inter-individual differences in NREM period length. On the relative time scale, t=0 designates the time at which the maximum in delta EEG power occurs for all participants, represented in the figure by dotted vertical lines. The top row displays the mean delta EEG power profile (natural-log-transformed) as a function of relative time. The middle row displays the mean normalized HF-HRV profile (square-root-transformed) as a function of relative time. The bottom row reveals the time-varying correlation between delta EEG power and HF-HRV; solid lines represent the estimated time-varying correlation functions, and the shaded areas represent point-wise 95% confidence intervals for the correlation functions. A correlation is deemed statistically significant if its 95% confidence interval does not include zero.

Figure 2

Comparison of time-varying correlations between delta EEG power and HF-HRV across different NREM cycles in the full sample. In each figure, the middle curve represents the estimated correlation difference between two NREM periods on the Fisher-transformed scale, and the upper and lower curves represent point-wise 95% confidence limits for the correlation difference. 95% confidence intervals on the Fisher-transformed scale that do not include zero represent a significant difference in correlation between NREM periods on the original correlation scale. The figures shown compare (a) NREM2 – NREM1, (b) NREM3 – NREM1, and (c) NREM3 – NREM2.

Figure 2

Comparison of time-varying correlations between delta EEG power and HF-HRV across different NREM cycles in the full sample. In each figure, the middle curve represents the estimated correlation difference between two NREM periods on the Fisher-transformed scale, and the upper and lower curves represent point-wise 95% confidence limits for the correlation difference. 95% confidence intervals on the Fisher-transformed scale that do not include zero represent a significant difference in correlation between NREM periods on the original correlation scale. The figures shown compare (a) NREM2 – NREM1, (b) NREM3 – NREM1, and (c) NREM3 – NREM2.

Figure 2

Comparison of time-varying correlations between delta EEG power and HF-HRV across different NREM cycles in the full sample. In each figure, the middle curve represents the estimated correlation difference between two NREM periods on the Fisher-transformed scale, and the upper and lower curves represent point-wise 95% confidence limits for the correlation difference. 95% confidence intervals on the Fisher-transformed scale that do not include zero represent a significant difference in correlation between NREM periods on the original correlation scale. The figures shown compare (a) NREM2 – NREM1, (b) NREM3 – NREM1, and (c) NREM3 – NREM2.

Figure 3

Delta EEG power, HF-HRV, and time-varying correlation for NREM-1, NREM-2 and NREM-3 in the control group of participants without clinically significant SDB or self-reported Insomnia (n=146). See Figure 1 legend for details.

Figure 4

Delta EEG power, HF-HRV, and time-varying correlation for NREM-1, NREM-2 and NREM-3 in the group of participants with clinically significant SDB (AHI > 15) (n=32). See Figure 1 legend for details.

Figure 5

Between-group comparison (SDB – Controls) of time-varying correlations between delta EEG power and HF-HRV for (a) NREM-1 and (b) NREM-3. In each figure, the middle curve represents the estimated difference in correlation between participants with clinically significant SDB and control participants on the Fisher-transformed scale, and the upper and lower curves represent point-wise 95% confidence limits for the correlation difference. 95% confidence intervals on the Fisher-transformed scale that do not include zero represent a significant difference in correlation between participants with clinically-significant SDB and control participants on the original correlation scale.

Figure 5

Between-group comparison (SDB – Controls) of time-varying correlations between delta EEG power and HF-HRV for (a) NREM-1 and (b) NREM-3. In each figure, the middle curve represents the estimated difference in correlation between participants with clinically significant SDB and control participants on the Fisher-transformed scale, and the upper and lower curves represent point-wise 95% confidence limits for the correlation difference. 95% confidence intervals on the Fisher-transformed scale that do not include zero represent a significant difference in correlation between participants with clinically-significant SDB and control participants on the original correlation scale.

Figure 6

Delta EEG power, HF-HRV, and time-varying correlation for NREM-1, NREM-2 and NREM-3 for participants with self-reported insomnia (n=25). See Figure 1 legend for details.

Figure 7

Between-group comparison (Insomnia – Controls) of time-varying correlations between delta EEG power and HF-HRV for (a) NREM-1 and (b) NREM-2. In each figure, the middle curve represents the estimated difference in correlation between participants with self-reported insomnia and control participants on the Fisher-transformed scale, and the upper and lower curves represent point-wise 95% confidence limits for the correlation difference. 95% confidence intervals on the Fisher-transformed scale that do not include zero represent a significant difference between participants with self-reported insomnia and control participants on the original correlation scale.

Figure 7

Between-group comparison (Insomnia – Controls) of time-varying correlations between delta EEG power and HF-HRV for (a) NREM-1 and (b) NREM-2. In each figure, the middle curve represents the estimated difference in correlation between participants with self-reported insomnia and control participants on the Fisher-transformed scale, and the upper and lower curves represent point-wise 95% confidence limits for the correlation difference. 95% confidence intervals on the Fisher-transformed scale that do not include zero represent a significant difference between participants with self-reported insomnia and control participants on the original correlation scale.