Deep sleep, slow waves and how much people record

Deep sleep is the stage scored as N3, also called slow wave sleep after the large slow oscillations that dominate the recording. It is defined by what the brain activity looks like, not by how rested a person feels on waking. ^[1] Many people look into these numbers after a wearable device reports that their nightly total was low. Before worrying about those scores, several measured facts from laboratory recordings clarify how deep sleep behaves across life.

Most deep sleep is lost by midlife

The sharpest change happens far earlier than most people expect. In 149 healthy men aged 16 to 83, the mean percentage of deep slow wave sleep fell from 18.9 percent in early adulthood, ages 16 to 25, to 3.4 percent by midlife, ages 36 to 50, and was replaced by lighter sleep rather than by waking. ^[3]

After midlife the pattern changes character. Between midlife and late life, ages 71 to 83, there was no further significant decrease in slow wave sleep, but time awake rose by 28 minutes per decade at the expense of lighter non-REM sleep. ^[3] Rather than draining away evenly across decades, slow wave sleep drops steeply during early adulthood and then remains relatively steady.

Lifespan meta-analysis and screening differences

The lifespan picture is consistent across a much larger sample. A meta-analysis of 65 studies covering 3577 people aged 5 to 102 found the percentage of slow wave sleep significantly negatively correlated with age, alongside falls in total sleep time and sleep efficiency and rises in the time spent awake after falling asleep. ^[2]

CIBDOL · Lifespan meta-analysis and screening differences

The same meta-analysis carries a warning that is usually dropped when its figures are quoted. The size of every age effect depended heavily on how carefully the studies had screened their participants, to the point of diminishing or masking the association altogether, so a single quoted percentage for a given age is less solid than it looks. ^[2] How researchers select participants shapes the recorded output substantially.

Scoring reliability and consumer trackers

Deep sleep is also among the stages trained humans agree on least. In a reliability programme in which more than 2500 scorers judged 1800 epochs, producing over 3.2 million scoring decisions, overall agreement was 82.6 percent, agreement on N3 was 67.4 percent and only N1 was lower at 63.0 percent, and telling N2 from N3 was singled out as particularly difficult. ^[1]

A consumer tracker does not record brain activity. Every figure above comes from polysomnography, which records the electroencephalogram, eye movement and muscle tone together, and even that is scored by humans who disagree about this stage a third of the time. Wearable devices estimate sleep stages indirectly using movement and pulse sensors rather than direct neural measurements.

What slow wave sleep does in the body

Slow wave sleep is associated with the consolidation of certain kinds of memory, through the coordination of slow oscillations, spindles and ripples described in the review literature. ^[4] Separately, in the same 149 healthy men, the amount of growth hormone secreted tracked the amount of slow wave sleep. ^[3]

The widely repeated idea that the brain washes itself during deep sleep rests on animal work. In mice, natural sleep and anaesthesia were both associated with a 60 percent increase in the space between cells and a faster clearance of beta-amyloid. That is a finding in mice and it has not been established the same way in people. ^[5]

Normal ranges rather than arbitrary targets

This article reports measured ranges from published recordings. It does not set a target for how much deep sleep anyone should get. Young adults average close to a fifth of their sleep in this stage, whereas adults in midlife average around a twentieth. Observed recordings reflect age and experimental methodology rather than a uniform standard to achieve.