Your neck and shoulder muscles hold your head, which weighs about as much as a bowling ball, and steady your arms all day. They are built for long, low effort, not for rest.

That makes them the first place a stress response shows up. Bracing is cheap for the body and useful in a real emergency, so the setting is easy to switch on.

It is small enough to miss. In a lab study of 62 women, mental arithmetic alone raised shoulder muscle activity, with nothing to lift and nothing to hold.

Nobody notices a rise that size while it is happening. What people notice is the afternoon, when the tissue that has been working since breakfast starts to complain.

The reason a mild brace hurts is time, not force. The smallest fibres are switched on first and off last, so they can stay busy for the whole meeting without a real break.

Recovery is also slower than the trigger. In one study the raised muscle activity was still there in a later, calmer task, which is why the ache outlasts the deadline.

Stress turns the volume up as well. After a stressful public-speaking task, healthy volunteers rated the same test sensations as more painful than before it.

So the loop is tension plus a more sensitive pain system plus worry about the pain. Calming the state around the muscle does more than attacking the muscle itself.

The big muscle across the tops of your shoulders and up the back of your neck runs from the base of your skull out to the shoulder blade. Its job is to hold your head up and keep your shoulder steady while your hands do something useful.

That is postural work, so it is designed for long stretches of very light effort. It is not designed for a full day of light effort with no gaps in it at all.

Recordings during ordinary computer work found that in some people a few of the smallest muscle units stayed active without a break for the whole half-hour session. The load was tiny. The lack of pauses was the notable part.

That is the shape of the problem in one sentence. Nothing here is heavy, and nothing here stops, and the tissue doing the work is the tissue that ends up sore.

Two routes carry stress into muscle. Nerves fire within a second of a threat, and stress hormones follow over minutes and keep the setting raised for longer.

You do not need anything physical to start either one. In a study of 62 women, doing mental arithmetic and a colour-word test raised shoulder muscle activity while they sat still, and their blood pressure and stress hormones rose alongside it.

A second experiment added the mental task on top of holding the arms out. Muscle activity rose about a fifth, and stayed raised into a later round of the same holding task once the mental task had finished.

That delayed settling matters more than the size of the rise. The body keeps the brace for a while after the reason for it has gone, which is exactly what people describe when the shoulders are still up at eight in the evening.

The usual objection is fair: I am not straining anything, so why does it ache? Because muscle soreness at this level is about duration, not about effort.

Muscle units are recruited in size order. The smallest go on first and come off last, so during hours of light postural work the same small fibres carry the load while the bigger ones stay clean and rested.

Full-day recordings of 118 women found that the people whose shoulder muscle almost never went quiet during work were about twice as likely to have persistent pain as the people whose muscle had frequent silent gaps.

So the useful target is not force, it is gaps. Anything that gives the muscle real pauses through the day is working with the mechanism rather than against it.

Desk work is the classic setting, and the evidence is more modest than the headlines. A review of computer work found only limited support for it causing neck and shoulder problems, with mouse work the most consistent thread.

Phones add their own version. Following 1,691 teenagers for a year, those who used a smartphone at the start were more likely to report neck and shoulder symptoms later, and the total hours did not predict it. How they used it did.

Sleep sits underneath all of it. In a Finnish cohort followed for two years, short or poor sleep at sixteen predicted neck pain at eighteen, and in 16,167 Swedish adults sleep trouble came with clearly higher odds of neck and shoulder pain.

Add the ordinary rest: age, previous injuries, heavy days, some medications and little activity. In that Swedish survey, brisk exercise went with less pain while gentle general activity did not, so stress is one thread in the rope, not the rope.

The best evidence here is training the sore muscle rather than resting it. In 48 women with long-standing shoulder muscle pain, ten weeks of specific strength work cut their worst pain by about four fifths.

The comparison group in that trial cycled with relaxed shoulders and got only a small, short-lived drop, so it was the local training doing the work rather than general fitness.

The dose is smaller than people expect. Thirty office workers doing two minutes a day with an elastic band for ten weeks reported around 40% less pain, and their muscle spent longer genuinely at rest during the working day.

In a larger trial of 447 office workers, one hour a week helped whether it was taken as one session or split into a few minutes most days. That is the most encouraging finding on this page, and the easiest one to act on this week.

One belief deserves retiring: that bad posture is the cause. When researchers photographed 1,108 teenagers and sorted them into four real posture types, the groups had the same odds of neck pain and headache as each other.

A second one goes with it. Stress does not squeeze the blood out of the muscle: when researchers measured flow inside the working muscle during a mental stress task, it went up, not down, even though the skin went pale.

The picture also contradicts itself in useful places. Comparing 18 women with long-term shoulder pain against 30 without, the pain group did not show the revved-up stress nervous system the theory predicted, and full-day recordings suggest the muscle-activity explanation only fits some people.

And the same long study that ties work stress to later pain found that leisure exercise did not offset frequent work stress. Hold the mechanism loosely, and treat the pattern as real but not fully mapped.