Stress: What the Body Actually Does When You Feel It

Stress is not the feeling of being overwhelmed. It is the body's nonspecific response to a demand, and it runs whether or not you notice it happening.

This article follows that response from the brain down to the adrenal gland and back up again, then follows what happens when it never gets switched off.

The advice you have read about stress usually stops at the part you can feel. The mechanism underneath it is where the real and the popular versions of stress advice come apart, so it is worth naming in detail.

Where the Word Came From

The original version of this post said the word stress was coined in the 1950s. That is two decades late.

Walter Cannon came first. Working at Harvard on the bodily changes that go with hunger, cold, injury and strong emotion, he gathered them into one response in 1915 and called it fighting-or-flight.

He also coined the word homeostasis a few years later, for the idea that the internal environment has to be held inside workable limits whatever the outside is doing.

Cannon is the source of the phrase most people know. The word stress is his successor's.

Hans Selye was a medical student at McGill in Montreal in 1936, trying to find new hormones in cow ovary extracts. He noticed that the same pattern of changes followed from substances that had nothing to do with ovaries.

Whatever he injected, the animals showed the same four things: an enlarged adrenal cortex, ulcers in the stomach and gut, a shrunken thymus, and shrunken lymph nodes.

That was the finding. Not a new hormone, but the observation that the body answers a demand the same way whatever the demand is, which is what the word nonspecific means here.

The three stages Selye described

From those experiments Selye built a three-stage sequence, and it is the sequence most people mean when they say stress.

The alarm stage is the acute reaction. It arrives fast, and it is the part you can feel.

The resistance stage follows, in which the acute signs settle down while the body holds on at a higher level.

The exhaustion stage is the collapse, where the first-stage reaction returns or the organism fails.

It is worth being honest about how that model has aged. The mechanism underneath it is now well described. The tidy three-step curve is a rodent finding, and it does not predict human chronic stress very well.

Why the physics metaphor

The post said stress was "rooted in physics" and left it there. That is right, and the specific debt is worth naming.

The idea was imported from Hooke's law, the 1660s statement that the restoring force in a stretched material is proportional to the stretch.

Stretch it and a force builds. That is the metaphor: a demand is a deformation, and the response is the restoring force that goes with it.

The word for the demand was "stress," and the word for the response was "strain." A bar under load is stressed. The body that answers the load is strained. Most people use the two words for the same thing, which is how the definition ends up describing a sensation.

Selye turned the metaphor into a biological definition in 1950: a nonspecific response of the body to any demand made upon it. That definition has no feeling in it, and that is deliberate.

The Cascade, Gland by Gland

Here is the whole chain. It is short, it is well mapped, and every step has been measured directly.

A group of neurons in the paraventricular nucleus of the hypothalamus starts it. They make corticotropin-releasing hormone, usually shortened to CRH.

Those neurons send CRH down to the median eminence, where it is released into a private set of veins called the hypophysial portal system.

This plumbing detail matters. The blood vessels carry CRH straight to the pituitary instead of around the whole body first, which is how the brain keeps this signal local.

CRH binds to receptors on cells in the front of the pituitary called corticotrophs. Those cells release adrenocorticotropic hormone, shortened to ACTH, into the bloodstream.

ACTH travels to the outer layer of the adrenal gland, the zona fasciculata, and binds to melanocortin 2 receptors there.

That binding is what makes the adrenal cortex build and release glucocorticoids. In people the main one is cortisol. In rats and mice it is corticosterone instead.

A dissected adrenal gland resting on the upper pole of a kidney. The thin outer rind is the adrenal cortex, the layer that releases cortisol in response to ACTH. LearnAnatomy, CC BY 3.0, via Wikimedia Commons

Cortisol then reaches receptors in nearly every organ system, including the brain. It is a wide signal, not a local one, which is why chronic exposure to it has effects everywhere at once.

A helper that cannot work alone

The CRH neurons do not release CRH by themselves. They also release vasopressin from the same nerve endings.

Vasopressin acts on a different receptor, AVP1B, and works through protein kinase C rather than the route CRH uses. It complements the CRH signal and strengthens it.

On its own, at normal levels, vasopressin cannot drive a meaningful ACTH response. It is a modulator that needs the main signal to be present.

That is a small detail that changes how you read the whole system. The output depends on the combination of what arrives, not on any one messenger alone.

Why one precursor makes two hormones

ACTH does not float into the blood on its own. It is cut out of a larger molecule, which is why the same factory makes something else too.

CRH switches on the gene for proopiomelanocortin, usually shortened to POMC. That single protein is the raw material.

An enzyme called prohormone convertase 1 cuts it in two. One piece is ACTH. The other is beta-lipotropin, which goes on to produce other peptides including the body's own opioids.

So the same cut that delivers the stress hormone also feeds a second system. Stress signaling in this axis is not one line with one product on the end.

A stained section through the adrenal gland, showing the layered structure of the cortex outside and the medulla inside. The zona fasciculata, the ACTH-responsive layer, is the broadest band. AmonNonza, CC BY-SA 4.0, via Wikimedia Commons

Two receptors, two jobs

Cortisol is one hormone. It works through two different receptors, and they behave in ways that sound wrong until you look at their affinity.

The mineralocorticoid receptor has the higher binding affinity. It is saturated by ordinary resting levels of the hormone, so it sets the background rhythm.

The glucocorticoid receptor has the lower affinity. It only gets properly occupied when levels climb into the range that stress produces.

That arrangement is why the axis has a baseline at all. You are not at zero between stressful events, and the reading you get in a quiet moment is not a reading of nothing happening.

The two receptors are not a switch and a dimmer. They are two different jobs, and the axis uses both at once.

Both are what scientists call ligand-gated transcription factors. Binding them changes which genes are read, and that is how a hormone lasting minutes produces changes lasting hours.

Higher levels push energy out of storage. They act on liver, fat and muscle, and they change how the immune system and the nervous system behave.

How it shuts off

A system that can be switched on and never off is a fault, not a design. So there is feedback, and it runs at several levels at once.

Cortisol acts in the brainstem to cut the drive going up. It acts on limbic structures such as the hippocampus to reinforce inhibition.

And it acts directly on the paraventricular nucleus, quickly, through membrane-bound glucocorticoid receptors, to silence the CRH neurons that started all of it.

Three nested brakes. Note that the fast one uses membrane receptors rather than the nuclear kind, so it is a different mechanism from the slow gene changes.

What Is Not a Stress Response

Two things get folded into stress that should not be, and getting them apart explains some results that otherwise look strange.

The first is the sympathetic nervous system. It is fast, it acts on seconds, and it releases adrenaline and noradrenaline.

The HPA axis is slower and hormone-driven. They are two different arms of one response, and only one of them is the cortisol axis.

The second is the daily rhythm. HPA activity rises at waking, falls across the day and bottoms out overnight.

Glucocorticoids are needed for energy balance even with no stressor at all. Liver, fat and muscle all need the signal to release stored fuel for ordinary use.

So a cortisol curve that rises after waking is not a curve of someone in distress. That is the normal shape of the axis doing its ordinary work.

Stress is best understood as an extra charge on top of a running system, not as the running system.

Two kinds of stressor, two pathways

Not every threat reaches the hypothalamus by the same road, and the difference is more than academic.

A reactive stressor is something happening now. Hunger, cold, blood loss, a physical blow. Sensory relays drive the paraventricular neurons directly, through noradrenergic and peptide signals.

An anticipatory stressor is something expected. A deadline, a diagnosis, a meeting you are dreading. No sensory input demands anything.

Anticipatory responses run through a different set of circuits, out of limbic structures upstream.

They work largely by disinhibition: the amygdala silences the tonic GABAergic brake on the paraventricular nucleus, and the neurons fire because nothing is holding them down.

This is why an imagined deadline produces a genuine hormonal response. Nothing is touching the body, and the axis does not care, because it is responding to the prediction rather than the event.

Why Chronic Is a Different Problem

An acute stress response that switches off is not a disease. The damage is in the failure to switch off, and in the response switching on too often.

McEwen and Gianaros named the useful distinction. Allostasis is the adaptive half: parameters shift so the body can meet a demand, then shift back.

Allostatic load is what that shifting costs. It is the wear and tear accumulated on the body and the brain when the shifting runs too often or runs too long.

They also sort the ways load builds into four, and the four are worth knowing because they look identical from the outside.

The four routes to load

The first is repeated hits. Many small stressors, over years, each one producing a real surge in blood pressure and cortisol activity.

The second is failure to habituate. The response stops falling with repetition. It was first described in people who kept mounting a cortisol response to a public-speaking task they had already given many talks on.

The third is a response that is slow to end. Blood pressure stays up across a repetitive, deadline-driven workday because the system that should be closing the response has not.

The fourth is an inadequate response, which is the least obvious one. When the glucocorticoid answer is too weak, other mediators compensate by going into overdrive.

Autoimmune conditions and chronic inflammation are where this one shows up, because inflammation is normally shut down by the glucocorticoid signal.

Good, tolerable and toxic

The same reviews sort stress experience into three kinds, and the sorting key is not how hard the event was.

It is how much control you have over it, and what support and resources are available to meet it.

Stress that is tolerable can produce growth, better adaptation and useful learning. Stress that is toxic does the opposite, and it raises the odds of illness and earlier death.

A demanding event with control and support behind it can be protective. A milder event with neither can be damaging.

The brain is also not a bystander here. It decides what counts as threatening, it runs the response, and it changes structurally in the process.

Three regions carry most of that work: the hippocampus, the amygdala, and the prefrontal cortex.

A classic anatomical plate of the base of the brain. The pituitary sits in its socket beneath the hypothalamus, which is the short physical link the whole HPA cascade depends on. Henry Vandyke Carter, public domain, via Wikimedia Commons

When the Hormone Stops Being Heard

Here is the finding that changed how chronic stress is explained, and it contradicts the most common story about cortisol.

The common story is that chronic stress floods the body with cortisol and the excess does the damage. The evidence points somewhere less obvious.

Cohen and colleagues argued that the hormone level may be the wrong variable. What may matter is how the target tissue responds to the hormone.

They call the failure glucocorticoid receptor resistance. Immune cells become less sensitive to the signal that normally terminates an inflammatory response.

So cortisol is still arriving. The cells have become partly deaf to it. The shutdown that should follow an infection does not happen on schedule or at the right strength.

Two studies made this concrete, both using deliberate rhinovirus exposure on volunteers in quarantine.

In the first, 276 healthy adults were assessed for recent long-term threatening stress and for receptor resistance, then exposed to one of two cold viruses and followed for five days.

People with high recent stress showed the resistance, and people who showed it were at higher risk of going on to develop a cold after being exposed to the virus.

The second study used 79 subjects and measured the local inflammatory response directly, in nasal secretions, after the same kind of exposure.

Greater resistance predicted more of the inflammatory signaling cytokines in the infected nose. The brake on inflammation was slipping.

One thing this does not show is that cortisol levels were low. Wake-up saliva and 24-hour urine measures did not account for the effect.

So the mechanism is not an overabundance of hormone. It is a loss of sensitivity in the tissue that is supposed to respond to it, which is a very different animal.

It also explains why chronic stress tracks with cardiovascular disease, type 2 diabetes and autoimmune conditions. Inflammation sits inside all three, and glucocorticoids are how the body normally ends it.

What chronic load does to the brain

The structural side of the story is measurable in large cohorts, and it is worth taking seriously rather than abstractly.

McManus and colleagues used UK Biobank data from 502,520 participants, aged 40 to 69. They split them into four groups by high and low stress in childhood and adulthood.

Higher childhood stress in women went with reduced connectivity in the posterior thalamic radiation and in the cingulum of the hippocampus. In men, the same kind of change came with higher adult stress.

High stress at both ages went with lower executive function and worse working memory in both sexes, and with more diagnosed mental health problems.

The mental health effect was stronger in women than in men, which matters given that severe childhood stressors are also more common in women.

This is a cross-sectional study, so it shows association rather than cause. That is a real limit and worth stating.

What it does establish is that this load has a measurable correlate in living people, measured on the timescale of decades.

A labelled diagram of the hypothalamus. The paraventricular nucleus, named near the top, holds the CRH neurons that start this cascade and the ones the cortisol brakes act on. Amparo Guemes and Pantelis Georgiou, CC BY 4.0, via Wikimedia Commons

Key Terms in Plain English

These are the words in this article that would send you to a dictionary. Each one is given here the way it is actually used above.

Term	What it means here
Stressor	Anything that makes a real or perceived demand on the body.
HPA axis	The hormone route from hypothalamus to pituitary to adrenal cortex.
CRH	The hypothalamic hormone that starts the HPA cascade.
ACTH	The pituitary hormone that tells the adrenal cortex to work.
Cortisol	The main glucocorticoid in people, released by the adrenal cortex.
Zona fasciculata	The adrenal layer that responds to ACTH.
Glucocorticoid receptor	The lower-affinity receptor that cortisol uses at stress levels.
Negative feedback	The set of brakes that limit how long the response runs.
Allostasis	Shifting internal set points to meet a demand, then shifting back.
Allostatic load	Accumulated damage from shifting set points too often or for too long.
Reactive stressor	A threat happening now, arriving through the senses.
Anticipatory stressor	A threat expected, driven by limbic circuits instead.

The words in this article that would otherwise send you to a dictionary.

Key Takeaways

• Selye published the stress syndrome in 1936, not the 1950s, and set out the biological definition in 1950.
• Stress is a nonspecific response to a demand. It is not the feeling of being overwhelmed.
• The physics source is real. The idea came from Hooke's law, which is why demand is stress and the response is strain.
• The chain is hypothalamus to pituitary to adrenal cortex, carried by CRH, ACTH and cortisol in that order.
• A private set of portal veins carries CRH straight to the pituitary, so the brain's signal stays local.
• Cortisol works through two receptors. The high-affinity one sets the daily rhythm; the low-affinity one handles stress levels.
• Three separate brakes shut the response down, and the fastest one uses membrane receptors rather than gene changes.
• The axis runs on a daily rhythm even without stress, so a morning rise in cortisol is normal.
• Imagined threats work because anticipatory responses disinhibit the hypothalamus through limbic circuits.
• Chronic load builds by four routes: repeated hits, failure to habituate, slow shutdown, and a response too weak to control inflammation.
• Whether a stressor is good or toxic depends on control and support, not on how severe it looks.
• Target cells can lose sensitivity while the hormone is still arriving in normal amounts.

Frequently Asked Questions

Is a morning cortisol rise a sign of stress?

No. HPA activity peaks at waking in most species and falls across the day. A rise after waking is the ordinary rhythm of an axis that is working as designed.

Does the HPA axis cause the fight-or-flight response?

No. The fast response is the sympathetic-adreno-medullary arm, acting on seconds. The HPA axis is the slower hormone arm. They are separate systems that get recruited together.

Does cortisol damage the body in chronic stress?

Not simply by being present. The receptor-resistance work suggests the harder problem is that target cells lose sensitivity, so a normal shutdown does not arrive on time.

Is Selye's three-stage model still right?

The mechanisms it implied are now well mapped. The tidy sequence is a rodent finding, and chronic human stress does not follow it cleanly.

Where does mindfulness or exercise fit?

The reviews used here do not test specific interventions. What they do show is that lifestyle behaviors both express allostatic load and add to it, and that control and support change how a stressor lands.

Can stress build up in childhood and show up decades later?

Yes. In a UK Biobank study of 502,520 participants, childhood and adulthood stress were both linked to microstructural changes in the hippocampus and to lower executive function and working memory.

Key references
• Herman et al. 2016, the stress response
• McEwen and Gianaros 2011, allostasis
• Cohen et al. 2012, receptor resistance
• Godoy et al. 2018, stress neurobiology
• McManus et al. 2022, stress across the lifespan

Images: Wikimedia Commons, with the author and license named in each caption. Figures, dates and study sizes follow the five sources above.