The Psychology of Stress: How Cortisol Reshapes the Brain

When physiologist Hans Selye accidentally injected ovarian extract into lab rats in 1936 and then, as a control, injected simple saline solution and got the same results — enlarged adrenal glands, shrunken immune tissue, gastric ulcers — he realized something important: the rats' bodies were responding to the ordeal of being injected, not to whatever was in the syringe. What he had discovered was a generalized biological response to demand, which he would go on to call stress. Nearly ninety years of research since then has mapped the molecular machinery of that response in remarkable detail — and revealed that while the system is extraordinarily well designed for short emergencies, running it continuously for months or years produces structural damage to the brain that is visible on MRI scans.

What Stress Actually Is: The Biology Before the Psychology

Stress is not an emotion. It is a coordinated physiological state — a set of cascading hormonal and neural changes that mobilize the body's resources for rapid response to a perceived threat or demand. The psychological experience of feeling stressed is downstream of this biological cascade, not its cause.

Stress response — the coordinated activation of the autonomic nervous system and the HPA axis in response to a real or perceived threat, producing physiological changes that enhance short-term survival capacity at the cost of long-term homeostasis.

Stressor — any stimulus — physical, social, psychological, or anticipatory — that activates the stress response. Stressors do not have to be objectively dangerous; they only need to be perceived as threatening or uncontrollable.

The distinction between acute and chronic stress is not just a matter of duration. It reflects two fundamentally different biological modes. Acute stress (the kind that lasts minutes to hours) triggers the sympatho-adrenomedullary (SAM) axis — producing the immediate adrenaline surge associated with fight-or-flight — and then the slower HPA axis. Chronic stress (persisting over weeks or months) involves persistent HPA activation with prolonged elevated cortisol levels, a qualitatively different physiological state with distinct, often damaging consequences.

Understanding stress means understanding that the same system that saves your life when a car runs a red light in front of you can, if it never fully deactivates, quietly erode your capacity for memory, rational decision-making, and emotional regulation.

Source: Selye, H. (1936), "A Syndrome Produced by Diverse Nocuous Agents," Nature, 138, 32 — the foundational paper establishing the concept of the generalized stress response.

Source: McEwen, B.S. (2007), "Physiology and Neurobiology of Stress and Adaptation: Central Role of the Brain," Physiological Reviews, 87(3), 873–904 — comprehensive review of the brain's role in stress response and allostatic load.

The HPA Axis: The Stress Hormone System in Detail

The hypothalamic-pituitary-adrenal (HPA) axis is the body's primary hormonal stress response system. It operates on a timescale of minutes to hours — slower than the immediate adrenaline surge but far more consequential for long-term health.

HPA axis — a neuroendocrine feedback loop connecting the hypothalamus, the pituitary gland, and the adrenal glands, which collectively regulate the production and circulation of cortisol in response to stress.

When the brain perceives a threat — through sensory input, memory, anticipation, or even a thought — the hypothalamus releases corticotropin-releasing hormone (CRH). CRH signals the pituitary gland, which releases adrenocorticotropic hormone (ACTH) into the bloodstream. ACTH travels to the adrenal glands, which sit atop the kidneys, and triggers the release of cortisol from the adrenal cortex.

This three-relay system takes roughly 15–20 minutes from initial threat detection to peak cortisol levels in the bloodstream. The delay matters: by the time cortisol peaks, the immediate danger may have passed, but the body is now primed for sustained high-alert functioning — elevated blood glucose, suppressed immune function, enhanced cardiovascular output, and altered brain activity patterns.

Under normal conditions, cortisol itself shuts the HPA axis down through negative feedback: cortisol receptors in the hippocampus and hypothalamus detect elevated levels and send a "stand down" signal that curtails further CRH release. This feedback loop is the system's off-switch. Chronic stress gradually impairs this feedback — one of the key reasons long-term stress is self-perpetuating.

The HPA axis is not merely reactive. It operates on a circadian rhythm, peaking in the morning (which contributes to morning alertness and prepares the body for daily demands) and bottoming out around midnight. Chronic stress can flatten this rhythm, eliminating the morning peak and keeping cortisol at a low but constant elevated baseline throughout the day and night.

Source: Ulrich-Lai, Y.M. & Herman, J.P. (2009), "Neural Regulation of Endocrine and Autonomic Stress Responses," Nature Reviews Neuroscience, 10(6), 397–409 — detailed review of HPA axis circuitry and feedback mechanisms.

Source: Lightman, S.L. (2008), "The Neuroendocrinology of Stress: A Never Ending Story," Journal of Neuroendocrinology, 20(6), 880–884 — review of HPA axis dynamics, ultradian rhythms, and pulsatile cortisol secretion.

Cortisol: What It Does and Why It Matters

Cortisol is often described simply as "the stress hormone," a label that misrepresents its complexity. Cortisol is essential for life — people with Addison's disease, which destroys the adrenal cortex, cannot survive without cortisol replacement therapy. The problem is not cortisol itself but the dose, duration, and timing of its presence.

Cortisol — a glucocorticoid hormone produced by the adrenal cortex, essential for regulating metabolism, immune function, and the body's response to stress. At appropriate levels and timings, it is adaptive; chronically elevated, it has neurotoxic effects on the hippocampus and prefrontal cortex.

In the short term, cortisol does exactly what a stress response requires. It mobilizes glucose from the liver, providing immediate fuel for muscle activity. It suppresses non-essential functions — digestion, reproduction, tissue repair, immune surveillance — that can wait until the emergency is over. It enhances attention and consolidates emotionally significant memories (which is why stressful events are often remembered vividly, at least initially).

The metabolic effects of cortisol are particularly important. Cortisol causes the liver to increase blood glucose through gluconeogenesis — manufacturing new glucose from amino acids and fats. Under acute stress, this is adaptive: the muscles need fuel. Under chronic stress, persistently elevated blood glucose contributes to insulin resistance, visceral fat deposition, and the metabolic syndrome that cardiovascular researchers associate with stress-related disease risk.

The immune effects are equally significant. Cortisol is a potent anti-inflammatory agent. In acute doses, this is useful — it prevents the immune system from overreacting during an injury. Under chronic elevation, immune suppression becomes a liability: wound healing slows, susceptibility to infection increases, and certain immune-mediated diseases worsen.

For the brain, the story is particularly consequential, and is covered in detail below.

Effect	Acute Cortisol (Hours)	Chronic Cortisol (Weeks/Months)
Blood glucose	Rises (adaptive — fuels muscles)	Chronically elevated (insulin resistance risk)
Immune function	Suppressed (prevents over-reaction)	Chronically compromised
Memory	Enhanced for emotionally significant material	Impaired, particularly explicit/declarative
Attention	Focused and narrowed	Hypervigilant but distractible
Hippocampal volume	No significant change	Measurable shrinkage (dendritic retraction)
Prefrontal cortex	Mildly impaired under acute stress	Structurally remodeled; decision-making impaired
Amygdala reactivity	Heightened	Heightened and potentially enlarged
HPA feedback	Normal	Impaired; cortisol fails to shut axis down

Source: Sapolsky, R.M. (1996), "Why Stress Is Bad for Your Brain," Science, 273(5276), 749–750 — landmark summary of glucocorticoid neurotoxicity research demonstrating hippocampal damage under chronic stress.

The Yerkes-Dodson Curve: When Stress Helps and When It Hurts

In 1908, psychologists Robert Yerkes and John Dodson published an experiment using electric shocks to motivate mice through a discrimination task. What they found — that moderate levels of arousal produced the best performance, with both too little and too much arousal impairing it — became one of the most cited findings in psychology, the inverted U-curve that still appears in every introductory psychology textbook.

Yerkes-Dodson law — the empirical relationship between arousal and performance, describing an inverted U-shaped curve in which moderate arousal produces peak performance on complex tasks, while both under-arousal and over-arousal impair it. The optimal arousal point is lower for complex tasks than for simple ones.

Think of it like engine RPM on a car. At idle, the engine can barely move the vehicle. At moderate RPM, it runs efficiently and powerfully. At the redline, it risks catastrophic failure. Stress functions as the cognitive throttle — the right amount produces focus, energy, and sharp thinking; too much produces the cognitive equivalent of engine failure.

The shape of the curve depends critically on task complexity. For simple, well-practiced tasks (hitting a baseball, doing arithmetic you've done thousands of times), the optimal arousal point is relatively high — extra stress can be motivating. For complex tasks (writing a nuanced analysis, solving novel problems, navigating a difficult social negotiation), the optimal arousal point is lower — moderate stress helps, but high stress shifts processing away from the prefrontal cortex (complex cognition) toward more reactive, automatic systems.

This is why athletes can perform physical skills under intense pressure that would impair a scientist trying to think through a complex experiment. The tasks have fundamentally different optimal arousal profiles.

The practical implication for cognitive performance is that the relationship between deadline pressure and output quality follows a curve, not a line. A deadline several weeks away produces too little arousal for many people; a deadline in five minutes produces too much for complex cognitive work. The "productive urgency" zone sits in between, and its exact location varies by individual, task type, and baseline anxiety level.

Source: Yerkes, R.M. & Dodson, J.D. (1908), "The Relation of Strength of Stimulus to Rapidity of Habit-Formation," Journal of Comparative Neurology and Psychology, 18(5), 459–482 — the original experiment establishing the inverted U relationship between arousal and learning performance.

Source: Diamond, D.M., Campbell, A.M., Park, C.R., Halonen, J., & Zoladz, P.R. (2007), "The Temporal Dynamics Model of Emotional Memory Processing," Neural Plasticity, 2007 — modern reappraisal of the Yerkes-Dodson curve integrating neuroscience and memory research.

How Chronic Stress Physically Remodels the Brain

This is where the science moves from the conceptually familiar to the genuinely alarming. Chronic stress does not merely make you feel bad — it causes measurable, structural changes in the brain, visible on MRI scans and confirmed by post-mortem histology.

The hippocampus — the region critical for forming and retrieving declarative memories and for spatial navigation — is particularly vulnerable to glucocorticoid toxicity. It has the highest density of cortisol receptors of any brain region, which is adaptive in the short term (it uses cortisol feedback to regulate HPA activity) but becomes a liability under chronic exposure.

Prolonged elevation of cortisol produces dendritic retraction in hippocampal neurons — the dendrites (the receiving branches of neurons) physically shrink and lose synaptic connections. CA3 pyramidal neurons, particularly involved in memory consolidation, are especially affected. At extreme and sustained levels, glucocorticoids can trigger apoptosis — programmed neuronal death.

Bruce McEwen's research at Rockefeller University demonstrated these structural changes in rodents under chronic restraint stress. Human neuroimaging studies — particularly work on individuals with major depressive disorder, PTSD, and Cushing's syndrome (which causes pathologically elevated cortisol) — have consistently found reduced hippocampal volume compared to controls. A meta-analysis by Videbech and Ravnkilde (2004) published in American Journal of Psychiatry reviewed 30 neuroimaging studies and found robust reductions in hippocampal volume in depressed patients, with the effect size larger in people who had experienced more depressive episodes.

The prefrontal cortex (PFC) — responsible for working memory, planning, impulse control, and rational decision-making — is also remodeled by chronic stress. Repeated stress causes retraction of dendritic arbors in the medial PFC and strengthens connectivity in the amygdala. The functional result is a shift in the balance of brain control: the rational, flexible prefrontal system is weakened, while the reactive, threat-focused amygdala becomes more influential. Under chronic stress, people become more reflexive and less reflective — more likely to react with fear or anger and less capable of the deliberate reasoning that might otherwise moderate those reactions.

Remarkably, some of these structural changes may be reversible. Animal studies have shown dendritic regrowth in the hippocampus and PFC after stress ends, particularly in enriched environments and with exercise. Human studies on antidepressant treatment — some classes of which promote neurogenesis in the hippocampus — support the possibility of structural recovery.

Source: McEwen, B.S. & Magarinos, A.M. (2001), "Stress and Hippocampal Plasticity: Implications for the Pathophysiology of Affective Disorders," Human Psychopharmacology: Clinical and Experimental, 16(S1), S7–S19 — key review linking chronic stress-induced glucocorticoid exposure to hippocampal structural changes.

Source: Videbech, P. & Ravnkilde, B. (2004), "Hippocampal Volume and Depression: A Meta-Analysis of MRI Studies," American Journal of Psychiatry, 161(11), 1957–1966 — meta-analysis of 30 neuroimaging studies confirming hippocampal volume reduction in depression, consistent with chronic cortisol exposure.

Allostatic Load: The Cumulative Wear of Sustained Stress

The concept of homeostasis — the body returning to a stable baseline after perturbation — captures something real but misses an important dimension. The physiologist Peter Sterling and cardiologist Joseph Eyer introduced the concept of allostasis in 1988 to capture the active, predictive process by which the brain adjusts physiological parameters in anticipation of demands, not just in response to them. Bruce McEwen later developed the concept of allostatic load to quantify the cumulative physiological cost of these adjustments.

Allostatic load — the cumulative physiological "wear and tear" on body systems produced by repeated or chronic stress responses, quantified through biomarkers including cortisol levels, inflammatory markers, blood pressure, waist-hip ratio, and metabolic indicators.

The allostatic load model makes a critical distinction: stress responses are not merely reactions to events. The brain predicts demands and pre-adjusts physiology accordingly — which is why anticipatory anxiety produces many of the same physiological changes as actually encountering the feared event. A student who dreads an exam starts activating their HPA axis days before the test, accumulating physiological cost from a threat that has not yet materialized.

Chronically high allostatic load — measured through batteries of biomarkers — predicts accelerated cognitive aging, elevated cardiovascular risk, immune dysfunction, and increased vulnerability to mood disorders. Research by McEwen and Stellar (1993) documented that allostatic load, not any single stressor, best predicted health outcomes over time.

The social determinants of allostatic load are stark. Chronic exposure to poverty, discrimination, neighborhood violence, or caregiving without support produces elevated allostatic load across populations. The concept helps explain why social inequality has direct biological consequences — it is not merely that stressed people make worse health choices but that sustained social stressors produce the same neurobiological damage as sustained psychological ones.

Source: McEwen, B.S. & Stellar, E. (1993), "Stress and the Individual: Mechanisms Leading to Disease," Archives of Internal Medicine, 153(18), 2093–2101 — introduced the allostatic load model and its application to stress-related disease risk.

Psychological Modifiers: What Makes the Same Event More or Less Stressful

Two people can experience the same objective event — a failed job interview, a difficult conversation with a supervisor, a near-miss car accident — and show dramatically different physiological stress responses. The stress is not simply "out there" in the world; it is constructed by the brain based on how the event is appraised.

Richard Lazarus's cognitive appraisal theory, developed through the 1960s and 1970s, proposed that stress results from an appraisal of a situation as threatening and exceeding available coping resources. The appraisal process has two components: primary appraisal (is this relevant and potentially threatening?) and secondary appraisal (do I have the resources to cope?). The same situation produces high stress when resources feel inadequate and low stress when resources feel sufficient.

Perceived control is one of the most powerful modifiers of the stress response. Research by Megan Gunnar and others has shown that when animals can predict and control an unpleasant stimulus — even if they choose not to — the HPA axis response is blunted compared to the same stimulus presented unpredictably and uncontrollably. The identical physical stressor produces different cortisol levels depending on perceived control.

Social support is the other primary buffer. A 2008 study by Eisenberger and colleagues found that the presence of a supportive partner reduced neural responses to social threat in the prefrontal cortex and subgenual anterior cingulate. Human proximity and social connection literally dampen the HPA axis response to threat — one of the cleaner demonstrations that psychological and social variables have direct neurobiological effects.

Interestingly, the nature of the support matters. Received support (being helped) is less consistently beneficial than perceived support (knowing help is available if needed). The mere knowledge that someone would help if asked appears sufficient to modulate threat appraisal, even without the support being activated.

Source: Lazarus, R.S. & Folkman, S. (1984), Stress, Appraisal, and Coping, Springer — the foundational theoretical account of stress as a transactional process between person and environment.

Source: Eisenberger, N.I., Taylor, S.E., Gable, S.L., Hilmert, C.J., & Lieberman, M.D. (2007), "Neural Pathways Link Social Support to Attenuated Neuroendocrine Stress Responses," NeuroImage, 35(4), 1601–1612 — demonstrated the neural mechanisms by which social support buffers the cortisol stress response.

Frequently Asked Questions

Q: What is the difference between acute and chronic stress physiologically?
A: Acute stress activates the SAM axis (immediate adrenaline response) and briefly activates the HPA axis (cortisol over 15–30 minutes), with both systems returning to baseline within hours. Chronic stress maintains persistent HPA activation, impairs the cortisol negative-feedback mechanism, alters circadian cortisol rhythms, and produces structural brain changes that acute stress does not.

Q: Can chronic stress actually shrink the brain?
A: In specific regions, yes. Chronic cortisol exposure causes dendritic retraction and, under extreme conditions, neuronal loss in the hippocampus and medial prefrontal cortex. Meta-analyses of neuroimaging studies consistently show reduced hippocampal volume in individuals with chronic stress-related conditions including depression and PTSD. Some recovery may occur when the stressor is removed, particularly in hippocampal regions that retain some capacity for adult neurogenesis.

Q: Why does stress impair memory if acute stress enhances it?
A: The effect is dose- and timing-dependent. Cortisol released shortly before or during encoding of emotionally significant material can enhance memory for that material by engaging the amygdala-hippocampus interaction. But sustained cortisol elevation over days to weeks impairs the hippocampal mechanisms needed for forming new memories and for recall of previously learned material. The same hormone that sharpens memory acutely becomes a toxin when chronically elevated.

Q: Is the Yerkes-Dodson curve scientifically valid today?
A: The original 1908 study is often oversimplified in textbooks — Yerkes and Dodson used electric shock intensity and simple discrimination tasks, not complex human performance. But the core concept of an inverted U relationship between arousal and performance on complex tasks has substantial support from modern neuroscience, particularly regarding prefrontal cortex function under high arousal. The curve is a useful model while acknowledging that the optimal arousal point varies by task type, individual, and context.

Q: Does exercise reduce chronic stress at the neurobiological level?
A: Yes. Aerobic exercise reduces basal cortisol levels, improves HPA axis feedback sensitivity, and promotes neurogenesis in the hippocampus — the same region most vulnerable to chronic stress damage. Research by Hillman, Erickson, and Kramer (2008) in Nature Reviews Neuroscience documented robust effects of physical activity on brain structure and cognitive function, including in stress-exposed populations.

Q: Why does perceived control reduce stress even when you do not use that control?
A: Control affects appraisal, not just outcomes. The knowledge that an unpleasant situation is escapable or controllable shifts the cognitive appraisal from "helpless" to "potentially manageable," which modulates the prefrontal cortex's regulatory input on the HPA axis. This is why giving people choices — even trivial ones — in stressful environments consistently reduces reported distress and cortisol levels.

For a connected look at how the brain processes threats at a shorter timescale, see The Psychology of Fear. The relationship between stress and decision-making connects to The Psychology of Decision-Making. To understand how habits form partly as automatic stress-reduction mechanisms, see The Psychology of Habits. The mechanisms of classical conditioning that shape stress-associated responses are examined in Pavlov and Classical Conditioning.

The Takeaway

Stress is a biological system optimized for a world where threats lasted seconds to minutes. In a world where stressors persist for months or years — financial precarity, dysfunctional relationships, relentless workload — the same system that once saved the lives of ancestors quietly degrades the brains of their descendants. The hippocampus shrinks, the prefrontal cortex retreats, the amygdala expands its influence, and decision-making tilts away from reflection and toward reaction. The science does not offer a simple fix, but it does provide a precise description of the mechanism — and with mechanism comes the possibility of intervention at the right points in the chain.

---

Read next:

• The Psychology of Fear
• The Psychology of Decision-Making
• Why We Feel Emotions: The Science