Reported sleep duration reveals segmentation of the adult life-course into three phases

Abstract

Classically the human life-course is characterized by youth, middle age and old age. A wide range of biological, health and cognitive functions vary across this life-course. Here, using reported sleep duration from 730,187 participants across 63 countries, we find three distinct phases in the adult human life-course: early adulthood (19-33yrs), mid-adulthood (34-53yrs), and late adulthood (54+yrs). They appear stable across culture, gender, education and other demographics. During the third phase, where self-reported sleep duration increases with age, cognitive performance, as measured by spatial navigation, was found to have an inverted u-shape relationship with reported sleep duration: optimal performance peaks at 7 hours reported sleep. World-wide self-reported sleep duration patterns are geographically clustered, and are associated with economy, culture, and latitude.

Fig. 1. Distribution of reported sleep duration.

a Distribution across 730,187 participants over 63 countries (M = 7.01 h, SD = 1.07 h). b Reported sleep duration is shorter for men (M = 6.94 h, SD = 1.04 h) than women (M = 7.07 h, SD = 1.09 h). c Reported sleep duration across the adult lifespan. This evolution can be split in 3 phases: a sharp decrease from 19 years to 33 years, a slower decrease from 34 years to 53 years, and a re-increase from 54 years onwards. d Proportion of short (5 h) and long (9–10 h) sleepers per gender and age group. e Reported sleep duration across age for each gender (381,153 men, 349,034 women), level of education (204,017 secondary, 526,170 tertiary), commute duration (291,822 less than 30 min, 254,362 30 min to 1 h, 183,764 h plus) and home environment (222097 city, 508090 non city). f Reported sleep duration across age for WEIRD (n = 526,136) and non-WEIRD (n = 204,051) populations. WEIRD stands for Western, Educated, Industrialized, Rich and Democratic. Data points correspond to the average reported sleep duration within 3-year windows. Error bars correspond to 95% confidence intervals.

Fig. 2. Association between reported sleep duration and spatial ability in the 3 identified age groups.

a Wayfinding task: spatial ability was quantified with the Sea Hero Quest (SHQ) app. Participants completed 2 training levels that did not require spatial ability, as the target was visible from the starting point. They also completed 4 wayfinding levels, where the participants were asked to memorize a map, and navigate as quickly as possible toward 3 checkpoints in a set order. b Training (motor ability) and Wayfinding (spatial ability) performance averaged for each sleep duration. See Table S3 for statistical analysis. Error bars correspond to 95% confidence intervals.

Fig. 3. Distribution of reported sleep duration across countries.

a Reported sleep durations averaged in 7 global clusters of countries defined by Maddison. In the first, second and third age groups, the sample sizes are, respectively, n = 321,406, n = 263,932 and n = 144,849. The samples sizes of the country clusters are available in Supplementary Fig 3. b Random intercepts clustered by countries from a linear-mixed model predicting reported sleep duration with age, age², gender, education, home environment, commute duration and their interaction with age as fixed effects and country as random effect. Colors reflect raw reported sleep duration (not corrected for fixed effects), as in panel c. c World map of reported sleep durations. d For each country, average reported sleep duration as a function of the absolute value of its average latitude (Pearson’s r = 0.52, p < 0.001). Error bars correspond to the standard errors.