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?
Homeostasis
- 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.
Allostasis
- 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.
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.
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.
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.
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.
Apply an episode
Secondary outcomes
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.
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.
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.
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.
Eight claims — true or false?
Predict each verdict before revealing it. These statements target the misconceptions that survive a first pass through the material.