Mind & Body · Wednesday, 26 August 2026

01 · Briefing · what happened

Your body wants a different bed temperature at 2am than at 5am, and a thermostat cannot know that

Mind & Body 1 min 32 sources

Your ability to regulate your own temperature changes as you move through sleep stages. A mattress that tracked those changes beat one held at a constant temperature - which means the best fixed setting is still wrong for part of every night.

30C / 33C

the two settings

REM and non-REM, on the adaptive mattress [1]

15.8 min

average morning grogginess

across 2,355 adults aged 19 to 92 [2]

+14.3 min

if you have anxiety

the largest single effect in that study [2]

30 min

less time awake at night

from exercise plus sleep education together [3]

At a glance

Body temperature is not one number you hold all night. Your ability to regulate it varies across sleep stages, which is why the same room can feel right at midnight and wrong at four. [1]
Researchers put 25 adults through three full nights of monitored sleep, measuring brain waves rather than asking how they felt. [1]
The three conditions were ordinary sleep, a mattress held at a constant 33C, and one that changed with the stage: 30C during REM sleep and 33C during non-REM. [1]
The changing mattress improved sleep quality measures against both of the others. [1]
The interesting comparison is not against ordinary sleep - it is against the constant 33C, which is a perfectly sensible setting, and still wrong for part of every night. [1]
This is why bedroom-temperature advice is so unsatisfying: any single number is a compromise across states with different requirements.
The same logic sits under the grogginess after you wake. In a study of 2,355 Korean adults, average morning sleep inertia was 15.8 minutes. [2]
It was not evenly spread: people with anxiety reported 14.3 minutes longer than those without - by far the largest effect in the study. [2]
Sleep inertia was shorter in people who slept longer and in morning types, and did not differ significantly by age group. [2]
Fixing sleep is rarely one lever. In a trial, high-intensity circuit training plus sleep-health education cut time awake after falling asleep by about 30 minutes, more than either component alone. [3]
The exercise alone and the education alone each shortened time-to-fall-asleep by around two minutes. Together they changed something else entirely. [3]
And it is worth the effort: pooled evidence links improved sleep quality to meaningfully lower depressive symptoms. [4]
The pattern shows up in children too: sleep in six- to twelve-year-olds tracks with cognitive, emotional and behavioural outcomes, and in adults recovering from panic disorder, better sleep tracks with the return of decision-making. [8][9]
Sleep hygiene education helps in adults aged 50 to 80, and among young people in one Bay Area survey the factors affecting sleep were mostly the ordinary ones. [5][10]

Forces in play

One setting, many states High→

a constant 33C mattress lost to one that changed with the sleep stage [1]

Anxiety in the morning High→

14.3 extra minutes of grogginess, the biggest effect in a 2,355-person study [2]

Combined interventions Easing

exercise plus sleep education beat either alone by a wide margin on night-time waking [3]

Screens at night Building↑

night-time screen use is repeatedly linked to worse sleep across student and adolescent populations [6][7]

In play REM sleep — the stage where the temperature requirement changes A fixed thermostat — the sensible compromise that is wrong part of the night Anxiety — the strongest predictor of how long you feel rough on waking Polysomnography — brain-wave measurement, rather than asking how people slept

How it unfolded

Falling asleep the body cools; time-to-sleep is what most interventions move [3]
Non-REM the warmer setting, 33C, suits this stage [1]
REM the requirement changes, and the cooler 30C works better [1]
On waking grogginess averages about 16 minutes, and far longer with anxiety [2]

Where this points

The useful question is not what temperature your room should be but whether anything in your night is fixed that should not be - the covers you cannot get out from under, the thermostat that holds one number, the alarm that does not care which stage you are in.

Also today

9 more stories on this beat.

Two hours a week seems to be the strength-training sweet spot

A study found 90 to 120 minutes of strength training a week associated with lower risk of death from cardiovascular and neurological causes and from any cause. New guidance separately argues for keeping the programme simple. [11][12]

Why it matters — A survey of 1,000 adults found widespread uncertainty about which muscles an exercise targets and what good form is - which is an argument for simple over optimal. [13]
How you breathe changes what you can lift

In trained lifters, three breathing strategies during maximal bench press - inhaling, exhaling, and holding the breath - produced different performance and different heart-rate-variability responses. [14]

Why it matters — Breath-holding raises pressure inside the chest, which stabilises the spine and also loads the cardiovascular system. It is a trade, not a technique.
Strength training does something to mood as well

An updated meta-analysis found psychological benefits from strength exercise in people who are overweight or obese, separate from any change in weight. [15]

Why it matters — The benefit does not wait for the scale to move, which matters because the scale is the thing most likely to make somebody stop.
Chronic pain affects about one adult in five

Chronic pain affects roughly 20.9% of US adults, more than diabetes. A meta-analysis found single-session behavioural interventions - far less demanding than the usual eight-session course - can help, and a phase 3 trial tested a full-spectrum cannabis extract for chronic low back pain, which affects over half a billion people worldwide. [16][17][18] Policy work separately argues the pipeline for non-opioid treatments is the bottleneck, and light-based therapy has been reviewed across randomised trials. [20][21]

Why it matters — A treatment that takes one session rather than eight reaches people who cannot attend eight, which is a different kind of effectiveness from a bigger effect size. [17]
Sometimes the intervention is reading the prescriptions

A randomised pilot found that a structured medication review improved both pain management and quality of life in chronic pain patients. [19]

Why it matters — Nothing new was prescribed. Somebody looked properly at what was already being taken.
Fibre, and what actually lives in you

Large-scale work links specific gut micro-organisms to cardiometabolic health, a systematic review examined which dietary changes actually shift the microbiome, and clinicians keep returning to fibre as the practical lever. Stress and diet both track with gut composition. [22][23][24][25]

Why it matters — The microbiome is the area where the gap between what is genuinely established and what is being sold is widest.
Cortisol is not the villain it is sold as

Cortisol is the body's principal stress hormone and governs a wide range of ordinary processes; the concern in the research is dysregulation of the axis that controls it, particularly in relation to autoimmunity, rather than the hormone itself. [26][27]

Why it matters — Almost every product marketed on 'lowering cortisol' is selling against a hormone you need in order to wake up.
Breathing practices have a measurable mechanism

A clinical review set out the neurophysiological mechanisms behind breathing-based well-being practices, rather than treating them as unexplained. [28]

Why it matters — The interesting part is that the mechanism is now specific enough to argue about, which is what separates a practice from a belief.
Hearing and vision, together

Roughly 70% of home care recipients have hearing or vision impairment, and researchers co-designed a support intervention with older adults living in Australian home care rather than for them. Sensory ageing affects both systems, usually at once. [29][30]

Why it matters — Two impairments at once is not twice one impairment: losing both removes the compensation each would have provided for the other. Elsewhere: plain water is enough for ordinary hydration, and older adults describe habit as something rebuilt through life's changes rather than installed once. [31][32]

02 · Lesson · why it matters

Right on average, wrong nearly always

When what a system needs changes over time and the control holding it does not, the best fixed setting is a compromise that suits no moment exactly.

How it works

A system's requirement changes over time
But the control is a single fixed setting
So the setting is chosen to be right on average
Average is a state the system is rarely actually in
So the setting is slightly wrong most of the time
And the fix is not a better number, it is a changing one

The twist

The best constant setting is a compromise between states, not the right answer to any of them. When a requirement moves and the control does not, being right on average means being wrong nearly always - by a little.

Where you've seen this

Office heating

one temperature for a building that is full at eleven and empty at four

Medicine doses

the same tablet morning and night, for a body that processes it differently at each

Staffing a shop

the same number of people on the floor at the quietest and busiest hour

Speed limits

one number for a road that is a school run at eight and empty at midnight

The catch

Variable is not automatically better. A changing setting needs to know which state you are in, and getting that wrong is worse than a decent constant. The study worked because sleep stages are measurable; most household settings are not.

And the whole of it

None of this is a failure of anybody's judgement. The thermostat has one number because it can only have one number. The advice says 18 degrees because advice has to say something. Your body, meanwhile, is running a schedule nobody involved can see - which is true of almost everything we set once and then stop thinking about.

03 · Lab · your turn

One Number For A Night

Pick what your bed does across eight hours of shifting sleep stages, and find that even the best single temperature is wrong for a quarter of the night.

04 · Truth · what's really going on

Stripped of the framing

The finding is not that a heated mattress helps. It is that a sensible constant temperature lost to a changing one - which means the advice everyone gives, a single number for the bedroom, is wrong for part of every night by construction.

Why it lands — Sleep advice is sold as a list of fixed settings, because a fixed setting is something you can buy or follow. 'It depends which stage you are in, several times a night' is true and cannot be printed on a box, so it does not circulate.

Claimed

What people said. Not yet a fact.

The mattress study

Adaptive thermal regulation improved sleep architecture compared with a constant temperature. [1]

The study evaluates a product category, and the paper says so - it is framed as evaluating a novel mattress. That does not make the physiology wrong; it does mean the comparison chosen was one the product could win.
The sleep inertia study

Average morning grogginess is 15.8 minutes, and anxiety adds 14.3. [2]

The measure is a single self-reported question - 'how long does it take to clear your grogginess?' The authors flag this themselves as not capturing the full picture.
Strength-training coverage

90 to 120 minutes a week is a sweet spot for longevity. [11]

An observational association between training volume and death risk. People who lift for two hours a week differ from people who do not in many ways besides lifting.

Verified

What we could actually stand behind.

The temperature comparison was measured with polysomnography across three separate nights per person. [1]

How we checked — Brain-wave measurement rather than self-report, with each of 25 participants acting as their own control across all three conditions - which removes most of the ways this kind of study usually goes wrong.
Exercise plus sleep education reduced time awake after falling asleep by about 30 minutes, more than either alone. [3]

How we checked — A trial with a control group and confidence intervals reported for each arm, and the combined arm's advantage over each single arm reported separately.
Chronic pain affects about 20.9% of US adults, more than diabetes. [16]

How we checked — A prevalence figure repeated across clinical reviews, and the comparison with diabetes is what makes it legible rather than just large.

Nobody knows

Open questions — ours included.

Whether any of this survives outside a sleep lab.

Twenty-five people, three monitored nights each, in a controlled setting. It establishes that the requirement changes; it does not establish what a mattress in your house would do. [1]
Which way the anxiety and grogginess link runs.

People with anxiety report much longer sleep inertia. Whether anxiety lengthens it, or a rough hour after waking feeds anxiety, or both, is not answerable from a cross-sectional study. [2]
Whether the strength-training figure is a dose or a marker.

Two hours a week is associated with living longer. It is also associated with being the kind of person who can find two hours a week. [11]
What the microbiome findings mean for one person.

Large studies link specific organisms to cardiometabolic health and reviews describe which diets shift composition. None of that tells an individual what their own gut needs. [22][23]

Who gains

Whoever sells the adaptive mattress — The study evaluates a novel real-time temperature mattress against a constant one - a comparison designed around the product's own claim. [1]
The supplement industry — Cortisol is the body's principal stress hormone and the research concern is dysregulation of the axis controlling it, not the hormone being high. Products sold on 'lowering cortisol' trade on the simpler story. [26][27]
Anyone selling a single fixed rule — 'Keep your bedroom at X' and 'take this at bedtime' are sellable precisely because they do not vary, which is the property this study undercuts.

Who pays

People following the fixed advice — A constant setting is a compromise across states with different requirements, so following it correctly still means being slightly wrong for part of every night. [1]
People with anxiety — They report roughly 14 extra minutes of grogginess every morning, and almost no sleep advice acknowledges that the morning is a different problem for them. [2]
Anyone who cannot attend eight sessions — The standard behavioural course for chronic pain is around eight sessions. The single-session work matters because the eight-session version excludes people by design rather than by effectiveness. [17]

05 · Hope · carry this

The grogginess you feel on waking has an average length, and it is about sixteen minutes. It is a state your body passes through, not a verdict on how the day will go.