SEATTLE—A habitual sleep duration of seven hours is associated with peak cognitive performance for all age groups except the elderly, according to data presented at the 29th Annual Meeting of the Associated Professional Sleep Societies. Sleep durations less than seven hours and those more than seven hours are associated with worse performance on cognitive tasks. In addition, cognitive performance may decline faster for younger people than for older people after sleep durations longer than seven hours.

Many studies have suggested that short and long sleep durations correspond with poor cognitive performance in older age groups. Other research in teenagers has consistently demonstrated that shorter sleep durations and later bedtimes are linked with lower grades and risk-taking behaviors. Yet the data have not been totally consistent. One literature review, for example, indicated that the relationship between short sleep duration and cognitive dysfunction might be stronger among middle-aged people than among the elderly.

Measuring Performance With Lumosity
Anne Richards, MD, MPH, Assistant Professor of Psychiatry at the University of California, San Francisco, and colleagues examined a large sample of people who played brain-training games on the Internet to clarify the ways in which age and habitual sleep duration predict cognitive performance. The investigators hypothesized that shorter sleep duration would correlate with worse performance at all ages; that peak performance would occur at a higher sleep duration in younger, compared with older, participants; and that performance would degrade faster beyond that peak in older, compared with younger participants.

Dr. Richards and colleagues chose three games from Lumosity, an online brain-training program, for the study. One was a version of the one-back test, the second was a visuospatial memory task, and the third was an arithmetic task. The study included more than 512,000 first-time players of the Lumosity games between January 2012 and September 2013. Each participant played one to three games, and the investigators analyzed the games separately to maximize the sample sizes. Participants also provided information about age, gender, educational attainment, and typical nightly sleep duration.

The researchers separated the population into groups that each spanned an age range of 10 years. The youngest group included people between ages 15 and 24. To maximize sample size, the oldest group combined people between ages 75 and 89.

Approximately 64% of the population was female. The youngest group included more than 180,000 people. The size of each successive group decreased as age increased, and the oldest age group included more than 5,000 participants. The population’s mean sleep duration was approximately seven hours. The youngest age group reported a mean of 7.2 hours of sleep. Middle-aged individuals reported the least sleep: 6.7 hours.

Sleep Had Less Effect on the Elderly’s Performance
Age was the biggest predictor of cognitive performance. Education also was a strong predictor of performance, but the researchers controlled for its effects. Dr. Richards found a subtle but statistically significant relationship between sleep duration and cognitive performance.

Shorter sleep durations were associated with poorer performance in the population as a whole and among young and middle-aged individuals. Contrary to the investigators’ expectations, peak performance occurred at seven hours of sleep in the five youngest age groups. Sleep duration did not appear to affect cognitive performance in the oldest age groups, however.

For sleep durations up to seven hours, younger people had a steeper increase in performance for each additional hour of sleep than older people for all three games. For the one-back test, younger people also had a steeper decline in performance with each additional hour of sleep beyond seven hours, compared with older people. For the other two tests, the researchers found no statistically significant difference in change in performance for sleep durations beyond seven hours between the younger and the older people.

Excess Sleep May Entail a Cost
Among the data’s limitations are that the study was cross-sectional and included insufficient demographic information. In addition, the investigators did not consider health status to protect participants’ privacy. As a consequence, it is not possible to examine whether comorbidities affect the decline in cognitive performance observed with long sleep durations. Other limitations are that habitual sleep duration was measured by self-report and that Lumosity brain games are based on, but are not, standardized measures of cognitive function.

“These findings are intriguing because they indicate that perhaps there could be a cost to excess sleep, even in younger individuals,” said Dr. Richards. “It makes one wonder whether models of the homeostatic recovery process should incorporate a curvilinear relationship between sleep duration and optimal brain functioning.” Further research may be necessary to distinguish the effects of sleep duration in young people and those of circadian shifts. The study also suggests that neurologists should consider the Internet as “a potentially viable and effective strategy for studying this topic,” Dr. Richards concluded.

—Erik Greb

How Much Sleep Is Required for Peak Cognitive Performance?