The HPA Axis, in One Picture
Every cortisol number on this site traces back to three glands and a feedback loop: the hypothalamus orders, the pituitary relays, the adrenals deliver, and cortisol itself carries the stop signal. This page draws the loop once, then shows where it breaks — because every stress intervention you will read about is really an intervention on one node of this picture.
What the evidence supports
- The cascade is settled science: hypothalamic CRH drives pituitary ACTH, which drives adrenal cortisol, usually peaking 15–30 minutes after a stressor.
- Cortisol feeds back to the brain and pituitary to brake its own production — the loop's off-switch.
- The HPA works alongside a faster autonomic arm: adrenaline acts in seconds, cortisol in minutes; both respond to the same alarm.
What remains uncertain
- Why chronically stressed people can show opposite patterns — high cortisol in some conditions, low in others — is an active research question.
- How much of feedback sensitivity is inherited versus shaped by experience.
- Which node to target first for a given person — the field can describe the loop better than it can tune it.
Evidence last reviewed: August 15, 2026. Conclusions may change as new research is published.
brain to glands, and back
Three Glands, One Loop
The HPA axis is a three-station relay with a return wire. Station one: a cluster of neurons in the hypothalamus releases CRH — corticotropin-releasing hormone — into a private blood vessel that runs straight to station two, the pituitary gland at the base of the brain. The pituitary answers by releasing ACTH into the general circulation. Station three, the adrenal glands sitting on top of the kidneys, receives ACTH and releases cortisol into the bloodstream — where it reaches essentially every tissue in the body, including, crucially, the brain that started the whole chain. The circuit closes when cortisol docks at receptors in the hypothalamus, the pituitary, and the hippocampus and signals: enough. Production slows. That return wire — negative feedback — is what makes the system a loop rather than a fuse, and it is the part that breaks first under chronic strain.
Two speeds share the same alarm. The autonomic arm — adrenaline and noradrenaline released within seconds — is the body's first responder: heart rate, breathing, and blood pressure shift before you have finished appraising the threat. The hormonal arm, the HPA cascade, is the slower second wave: minutes to peak, an hour to clear. Together they are the stress response as a whole (Ulrich-Lai & Herman, Nature Reviews Neuroscience, 2009). The distinction matters practically: fast breathing exercises calm the autonomic arm within minutes, while the hormonal arm responds to rhythm and recovery over days — which is why the fixes for each live in different places on this site.
The Loop, Drawn
Read the diagram once and every stress story on this site becomes mechanical. An alarm — a threat, a deadline, the anticipation of either — is appraised by brain circuits that sit above the hypothalamus. CRH flows; ACTH follows; cortisol peaks fifteen to thirty minutes later, which is why the chemistry of stress is always slightly behind the feeling of it. And the dashed line is the whole game: cortisol reporting back to shut itself off. A loop with a working return wire self-corrects. A loop with a frayed return wire runs hot.
The Off-Switch: Negative Feedback
Cortisol's second job — after delivering energy — is turning off its own production line. It docks at glucocorticoid receptors in the hypothalamus, the pituitary, and the hippocampus, and the signal it sends is the same at all three stations: reduce output. Researchers can probe this brake directly with the dexamethasone test: give a synthetic cortisol, and in a healthy system the pituitary reads it as "enough" and stops releasing ACTH. In a subtype of severe depression, the brake fails — ACTH keeps flowing despite the signal, a finding that helped establish melancholic depression as a feedback disorder, not a mood metaphor (Carroll et al., Archives of General Psychiatry, 1981).
Receptor sensitivity, not just hormone levels, is where individual differences live. Two people can have identical cortisol numbers and different stress experiences because their receptors listen at different volumes — one of the reasons a single cortisol value means so little on its own. The same logic explains a paradox that confuses people: chronically stressed individuals can show low cortisol, not high, when the system has run the brake so hard it overshoots, or when receptors have become unusually sensitive. The loop's state is the story; the level is a snapshot of one moment in it.
Where the Loop Breaks
Every failure mode of this axis has a name and a signature:
- 🔇 The brake fails. Feedback resistance — the dexamethasone non-suppression above — appears in melancholic depression and after long glucocorticoid exposure (Carroll et al., 1981). The loop runs hot because the stop signal is ignored.
- 📉 The output sags. The opposite profile — blunted morning rise, low output — shows up in burnout, fatigue states, and post-traumatic stress. A meta-analysis of the chronic-stress literature found both directions: high cortisol in some conditions, low in others, with flattening across the board (Miller, Chen & Zhou, Psychological Bulletin, 2007). The axis dysregulates; it does not simply "run high."
- 🌊 The rhythm flattens. Chronic strain erodes the daily curve — the signature mapped on the rhythm page.
- 😴 Sleep removes the natural brake. One short night measurably lifts next-evening cortisol (Leproult et al., Sleep, 1997) — sleep is the loop's daily reset, and skipping it is how the loop quietly drifts hot.
- 🔁 The break point decides the symptoms. Brake failure looks like agitation and metabolic strain; sagging output looks like exhaustion and fog — which is why "high cortisol" is never the whole diagnosis.
Why the Picture Dictates the Fixes
Once you can see the loop, the site's stress interventions stop looking like a list of vibes and start looking like a map of nodes:
- 🫁 Slow breathing works on the fast arm. Extended exhales stimulate the vagal brake and calm the autonomic first responder within minutes — the Breathwork topic owns this node.
- 😴 Sleep and light work on the rhythm node. The daily reset restores the curve — the sleep protocol is, among other things, an HPA intervention.
- 🏃 Exercise uses the axis productively. A training session is a deliberate, recoverable activation — the spike-plus-recovery pattern that keeps the system responsive instead of stuck.
- 👥 Company works on the appraisal node. A threat faced with someone else registers differently before the cascade even starts — the Biology of Connection topic maps that chemistry.
- 🧘 Meditation works on the baseline. Repeated practice appears to lower resting arousal — the meditation protocol reviews the evidence honestly.
🔁 A loop, not a lever
You cannot dial cortisol like a thermostat, because the system dials itself — that is what the feedback wire is for. Every honest stress intervention works on a node of the loop (the alarm, the relay, the rhythm, the brake) rather than on the hormone directly. The practical consequence: aim for a loop that returns to baseline, not a lower number. The number will take care of itself.
A Minute-by-Minute Walkthrough
Put the picture in motion with a real event. Your phone buzzes: the client wants the presentation a day early. Seconds: brainstem and amygdala fire; adrenaline hits — heart rate up, palms damp. Within a minute or two: hypothalamic neurons release CRH; the pituitary relays ACTH. You feel the familiar edge but you are still chemically early in the wave. Ten to thirty minutes later, cortisol peaks: glucose up, attention narrowed, immune cells mobilized — you are now optimally resourced to re-plan the week. The task gets re-scoped, the threat resolves, and cortisol, meeting its own receptors, tells the chain to stand down. Within about an hour the wave has cleared, leaving the adaptation: the system practiced a full cycle, which is exactly what keeps it calibrated. Contrast the flat version — the worry that outlasts the re-scoping, the evening replay, the short sleep — and you have the difference between using the loop and idling it hot. The acute-vs-chronic page is that contrast, expanded.
Questions, Answered Briefly
- 🪫 Can my adrenal glands "run dry"? The popular adrenal-fatigue story is weakly supported in endocrinology: the glands rarely run out. What happens is dysregulation — the loop runs flat, hot, or out of rhythm, not empty.
- 🧪 Can I test my axis at home? A saliva curve approximates the output side of the loop (see the measuring page); the feedback side — the dexamethasone test — is clinical territory.
- 💥 Why do I crash after the deadline, not during? Cortisol peaks 15–30 minutes behind the event and clears over an hour. The chemistry lags the alarm — the crash after is the wave receding.
- 🌿 Do adaptogens repair the axis? Trials are small and short, and none show the loop actually resetting — the free fixes (sleep, recovery, breathing) have the better evidence. The Pitfalls topic audits the supplement shelf.
The Bottom Line
- The HPA axis is a relay with a return wire — CRH, ACTH, cortisol, and the negative feedback that turns it off.
- Negative feedback is the health of the system — a loop that returns to baseline is working; a loop that stays hot or flat is not.
- Dysregulation runs in both directions — high in some states, low in others — so a single cortisol number is rarely the diagnosis.
- Interventions map onto nodes — breathing, sleep, exercise, company, and meditation each act on a different part of the loop.
Related Topics
- Ulrich-Lai & Herman, "Neural regulation of endocrine and autonomic stress responses," Nature Reviews Neuroscience (2009)
- Herman et al., "Regulation of the hypothalamic-pituitary-adrenocortical stress response," Comprehensive Physiology (2016)
- de Kloet, Joëls & Holsboer, "Stress and the brain: from adaptation to disease," Nature Reviews Neuroscience (2005)
- Carroll et al., "A specific laboratory test for the diagnosis of melancholia: standardization, validation, and clinical utility," Archives of General Psychiatry (1981)
- Miller, Chen & Zhou, "If it goes up, must it come down? Chronic stress and the hypothalamic-pituitary-adrenal axis in humans," Psychological Bulletin (2007)
- Raison & Miller, "When not enough is too much: the role of insufficient glucocorticoid signaling in the pathophysiology of stress-related disorders," American Journal of Psychiatry (2003)
- Chrousos, "Stress and disorders of the stress system," Nature Reviews Endocrinology (2009)
- Leproult et al., "Sleep loss results in an elevation of cortisol levels the next evening," Sleep (1997)
- Weitzman et al., "Twenty-four hour pattern of the episodic secretion of cortisol in normal subjects," Journal of Clinical Endocrinology & Metabolism (1971)