Graduate Physiology · Interactive Tool

Stability, Two Ways

The body has two stories for how it stays alive. One corrects errors. The other predicts them. Knowing which is which reorganizes how you read chronic disease.

How does the body hold itself together — by correcting what has already gone wrong, or by predicting what is about to?
● 4 interactive tools ● ~25 min ● Homeostasis · Allostasis · Allostatic load
1
The Core Distinction

Two models, one deeper question

Homeostasis and allostasis are often taught as rivals. They are better understood as different answers to a deeper question: what is the body actually regulating, and when does it act?

Model A

Homeostasis

Walter Cannon, 1929 — extending Claude Bernard's milieu intérieur (1865)
Claim
Regulated variables are held within a narrow range around a fixed setpoint.
Mechanism
Negative feedback — a sensor detects deviation, an effector corrects it.
Timing
Reactive. The correction follows the error.
Signal
The error itself drives the response.
Metaphor: a thermostat. The furnace fires only after the room has already gone cold.
Model B

Allostasis

Sterling & Eyer, 1988 — formalized by Sterling (2012) & Bruce McEwen
Claim
Stability through change — parameters are adjusted in advance to meet anticipated demand.
Mechanism
Predictive (feedforward) regulation orchestrated by the brain.
Timing
Anticipatory. The adjustment precedes the demand.
Signal
Prediction — prior experience plus current context.
Metaphor: a smart building that pre-heats before occupants arrive — using the calendar, not the cold.
Dimension
Homeostasis
Allostasis
Goal
Hold each variable constant.
Keep the whole organism viable and efficient; constancy is only sometimes the means.
Setpoint
A fixed reference value.
Re-set continuously by the brain to match predicted need.
Timing
Reactive — corrects after an error appears.
Predictive — acts before the error occurs.
Scope
Local, single-variable feedback loops.
Global — the brain trades off across systems to minimize total cost.
What is regulated
The variable.
The prediction — and, ultimately, fitness.
The reframe

The brain defends a forecast, not a number

Sterling's central move: a "setpoint" that the brain shifts on a daily, hourly, or anticipatory schedule is not a setpoint — it is a prediction. Blood pressure rises before you stand. Cortisol rises before you wake. Insulin is released before glucose climbs.

Negative feedback cannot, even in principle, produce a response to an error that has not happened yet. Wherever physiology acts in advance, feedback is not the whole story.

Not rivals

Allostasis explains how the non-negotiables stay constant

Homeostasis is not wrong. A few variables — arterial pH, blood oxygen, core temperature, plasma calcium, brain glucose — have tolerances so narrow that error is catastrophic. These are genuinely defended near-constants.

Allostasis explains the mechanism: the brain holds those few non-negotiables steady precisely by letting everything else — heart rate, blood pressure, ventilation, hormone levels, behavior — vary widely and predictively. The negotiable variables are moved so the non-negotiable ones don't have to.

Try it — which regime governs each variable?

Click a variable to classify it.

A variable's regulatory regime — and the reason for it — will appear here.
An honest aside

Is allostasis just homeostasis with feedforward added?

Feedforward control was already known to Cannon, and some critics argue allostasis renames familiar physiology. Defenders answer that the reframing is substantive: it changes what counts as the regulated quantity (whole-organism viability, not isolated variables), relocates control to the brain, and reinterprets setpoints as predictions. Whether that is a paradigm shift or a useful relabeling is genuinely unsettled.

The clinical value of allostatic load, however, is far less contested — it predicts morbidity and mortality across many cohorts. That is the engine of the next section.

2
Interactive · Allostatic Load Simulator

The price of adaptation

Allostasis is the process of adaptation. Allostatic load is its price — the cumulative biological cost of activating mediators too often, for too long, without habituation, or without an adequate shut-off. McEwen identified four patterns that generate load. Apply them to a 30-day timeline and watch the cost accumulate.

Cortisol Inflammatory load Tolerance band
Allostatic load 0
Adaptive — no accumulated cost
Add an episode to begin. Start with an acute healthy stressor to see the reference case.

Apply an episode

Secondary outcomes

Systolic BP112 mmHg
HbA1c5.1 %
hs-CRP0.6 mg/L
What to notice

A response that turns off is not damaging

The acute healthy episode barely moves the load — it spikes and fully recovers. Load is the cost of the four failure patterns, and recovery genuinely reverses it. This is why effective intervention targets the dynamics of the response — sleep, habituation, shut-off, recovery — not only the count of stressors.

From model to measurement

The allostatic load index

In the clinic, load is operationalized as a composite. The MacArthur studies (Seeman et al., 1997) count how many of roughly ten biomarkers — systolic and diastolic blood pressure, waist-hip ratio, HDL and total/HDL cholesterol, HbA1c, urinary cortisol, norepinephrine and epinephrine, and serum DHEA-S — fall in the high-risk quartile. The composite predicts mortality, cardiovascular events, and cognitive and physical decline better than any single marker. A cluster of sub-threshold findings is the signal.

3
Interactive · Dual-Lens Scenarios

The same event, read two ways

Pick an everyday physiological event. Read it first through the homeostasis lens, then the allostasis lens. The point is not that one lens is wrong — it is to locate the anticipation gap: the part of the response that happens before any error exists, which feedback alone cannot explain.

Homeostasis lens — react

Allostasis lens — predict

The anticipation gap

4
Interactive · Clinical Cases

Reading pathophysiology through the load lens

Four graduate-level vignettes. Commit to an interpretation before revealing the analysis — the wrong answers are the instructive ones, because each encodes a common misreading.

5
Interactive · Concept Self-Check

Eight claims — true or false?

Predict each verdict before revealing it. These statements target the misconceptions that survive a first pass through the material.

0 of 8 worked through