Reference Data for Polysomnography-Measured and Subjective Sleep in Healthy Adults

Abstract

Study objectives: Reference data for sleep are needed for the interpretation of clinical sleep parameters. This analysis aimed to provide polysomnography-measured, spectral analytic and subjective reference data based on a sample of healthy adults. In addition, effects of age and sex were investigated.

Methods: The sample was selected from the archival database of the Sleep Center at the University Medical Center Freiburg and consisted of 206 healthy adults aged 19 to 73 years. For an adaptation and a second examination night, polysomnography parameters, spectral analytic data, and subjective sleep estimations are presented.

Results: With increasing age, sleep became shorter (less total sleep time, more wake time after sleep onset) and lighter (eg, more percentage of stage N1 sleep, increase in fast activity in the beta range). Sleep in females was deeper than sleep in males (eg, higher percentage of stage N3 sleep). Females had higher overall power density than males. Altogether, it is apparent that sleep parameters exhibit high standard deviations, suggesting a high variability within healthy adults and complicating the specification of reference values.

Conclusions: Our data suggest that the informative value of sleep reference data in healthy individuals is limited because of high interindividual and intraindividual variation within sleep variables. More research, preferably in the form of meta-analyses and/or large international databases, is needed to further investigate the relevance of such reference data for mental and physical health. In the absence of such knowledge, giving patients feedback about deviations from the norm in their sleep profile may give rise to ill-founded concerns and worry.

Keywords: healthy adults; polysomnography; reference data; sleep; subjective sleep.

Figure 1. Percentile curves for total sleep time, wake time after sleep onset, sleep onset latency, sleep efficiency, and percentages of wake and different sleep stages.

Figure 2. Percentile curves for logarithmic spectral power during stage N2 sleep.

Percentile curves for logarithmic spectral power during stage N2 sleep for alpha (8–12 Hz), beta 1 (16–24 Hz), beta 2 (24–32 Hz), delta 1 (0.1–1.0 Hz), delta 2 (1.0–3.5 Hz), theta (3.5–8 Hz), gamma (32–48 Hz), sigma 1 (12–14 Hz), and sigma 2 (14–16 Hz).

Figure 3. Percentile curves for logarithmic spectral power during stage R sleep.

Percentile curves for logarithmic spectral power during stage R sleep for alpha (8–12 Hz), beta 1 (16–24 Hz), beta 2 (24–32 Hz), delta 1 (0.1–1.0 Hz), delta 2 (1.0–3.5 Hz), theta (3.5–8 Hz), gamma (32–48 Hz), sigma 1 (12–14 Hz), and sigma 2 (14–16 Hz).

Figure 4. Sleep EEG spectral power density in stage N2 sleep and stage R sleep.

Bands were defined as follows: alpha = 8–12 Hz, beta 1 = 16–24 Hz, beta 2 = 24–32 Hz, delta 1 = 0.1–1.0 Hz, delta 2 = 1.0–3.5 Hz, gamma = 32–48 Hz, sigma 1 = 12–14 Hz, sigma 2 = 14–16 Hz, theta = 3.5–8 Hz.