Exercise for stress: how much, how hard, how soon it works

Exercise is the most over-recommended and under-specified stress advice there is. Everyone says to do it. Almost nobody says how much, how hard, or when it will start to matter. The trial evidence is better than the advice, so here it is in numbers.

A 2023 umbrella review in the British Journal of Sports Medicine pooled 97 systematic reviews covering more than 128,000 people. Physical activity reduced symptoms of depression, anxiety, and psychological distress across every population studied, and the effect sizes were in the same range as psychotherapy and medication in comparable trials. That is a strong result from a large body of evidence. The review also found the benefits were largest in people who started with the highest distress, which is the opposite of what many stressed people assume.

Not very. The same review found moderate and vigorous intensity produced similar mental health benefits. A brisk walk where you can talk but not sing counts. Running, cycling, swimming, weights, yoga, and dance all show effects; the type matters far less than doing it. For anxiety specifically, both aerobic work and resistance training reduce symptoms, and resistance training has slightly better evidence for people who dislike cardio.

The one intensity finding that matters: very hard sessions late in the evening can delay sleep. If you train after 8pm, keep it moderate.

The dose in most trials is around 150 minutes a week, which is the general health guideline. Three 50-minute sessions or five 30-minute ones both work. Shorter bouts add up: a 2022 analysis of 15 cohorts found that the largest drop in depression risk came from going from nothing to about half the recommended amount. The first hour a week does more than any later hour.

If you are starting from zero, ten minutes a day is a real dose, not a token one. Two hours outside: the walk that does the most covers the lowest-effort version.

Two different timescales, and it helps to know both.

Within the hour. A single session of moderate exercise lowers state anxiety for several hours afterwards. If you are wound up before a difficult afternoon, a 20-minute walk at lunch is not a wellness gesture; it is a measurable reduction in how the afternoon will feel.

Within six weeks. The trait effect, where your baseline anxiety and mood shift, builds over about four to eight weeks of consistent sessions. The trials that report week-by-week results tend to show separation from controls somewhere around week four. That matters because most people quit at week two, exactly when the acute benefit has become familiar and the baseline shift has not yet arrived.

Several things at once. Exercise raises brain-derived neurotrophic factor, which supports the hippocampus, a region that shrinks under chronic stress. It gives the stress response a job: the same cortisol and adrenaline that sit unused during a stressful desk day get spent. It improves sleep, and sleep improves everything else. And it gives the mind a stretch of time where the loop is interrupted by breathing and effort.

The obstacle is not knowledge. Stressed people know they should exercise and skip it because it feels like one more demand. Three things help:

• Tie it to something fixed. After the school run, before the first meeting, the walk home from the station. Decisions cost energy; routines do not.
• Lower the bar on bad days. Ten minutes counts. Keeping the habit alive matters more than the dose on any single day. Build a stress habit that survives your busiest weeks is about exactly this.
• Track mood, not distance. Note how you feel an hour after, for two weeks. Once you have seen the pattern in your own numbers, motivation stops being the problem.