Physiology Foundations: Gain, Sensitivity, and Response Speed in Control Systems

Gain: how strongly the system responds to an error

Gain describes the “strength” of a control response relative to the size of the error. If the error is large, a high-gain system produces a large corrective action; a low-gain system produces a small corrective action.

In physiology, gain is not a single organ property; it is an emergent property of the loop: sensor sensitivity, controller processing, effector strength, and delays all shape the effective gain you observe in the output.

A simple working definition you can calculate

For step-by-step numeric reasoning, use a simplified discrete-time rule:

new value = old value + (gain × error)

where:

• error = set point − current value
• gain is a number between 0 and 2 in this toy model (values > 1 often overshoot; values near 2 can oscillate)

This is not a full physiological model; it is a learning tool to connect gain to stability, overshoot, and oscillation.

Continue in our app.

• Listen to the audio with the screen off.
• Earn a certificate upon completion.
• Over 5000 courses for you to explore!

Or continue reading below...

Download the app

(1) Numeric examples: low vs. high gain

Example A: low gain → slow correction

Assume a set point of 100 units. The system is at 80 units, so the error is 20.

Choose gain = 0.2

Interpretation: The system approaches the set point gradually. It is typically stable and unlikely to overshoot, but it may be too slow to protect function when rapid correction is needed.

Example B: moderate gain → faster correction without overshoot

Choose gain = 0.7

Interpretation: Faster approach with diminishing corrections as the error shrinks. This is often the “sweet spot” behavior: quick enough, still stable.

Example C: high gain → overshoot

Choose gain = 1.3

Interpretation: The first correction is so strong that the output crosses the set point (overshoot). The system then reverses direction. If gain is high but not extreme, the oscillations may dampen and settle.

Example D: very high gain → sustained oscillation (especially with delays)

Choose gain = 2.0

Interpretation: The system flips back and forth around the set point. In real physiology, time delays (transport time, hormone kinetics, neural conduction, tissue response time) make oscillation more likely at lower gains than this toy example suggests.

Gain and stability: why “stronger correction” can become unstable

Stability depends on whether the corrective action reduces the error without repeatedly overcorrecting. High gain increases the risk of:

• Overshoot: the response crosses the set point because the correction is too large for the remaining error.
• Oscillation: repeated overshoot in alternating directions, often amplified by delays.
• Instability: oscillations that grow or fail to dampen, producing large swings in the controlled variable.

A practical way to think about it: gain sets how “aggressive” the system is; delays and inertia determine how long the system keeps changing after the command is issued. Aggressive commands plus slow-to-stop physiology is a recipe for overshoot.

(2) Graphs: slow correction, overshoot, oscillation

The following ASCII plots show the controlled variable over time after a disturbance. The dashed line is the set point.

Graph 1: low gain (slow, stable approach)

Value  ^                 _________ (approaches slowly)          set point = 100  ----  - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -  100|                         . . . . . . . . . . . . . . . . . . . . . . . . .  90|                   . . .  80|_____________.______________________________________________> time

Pattern: monotonic rise toward the set point, minimal risk of overshoot, longer response time.

Graph 2: higher gain (overshoot, then settle)

Value  ^                         /
                        /  \__ (damped) 100| - - - - - - - - - - - - - -/----\--------------------------  90|                    _/      \_  80|___________________/________________________________________> time

Pattern: crosses the set point, then returns with smaller swings (damping).

Graph 3: very high gain and/or significant delay (oscillation)

Value  ^                    /\    /\    /\    /\ 100| - - - - - - - - - - -/--\--/--\--/--\--/--\-------------  90|                  /    \/    \/    \/    \  80|_________________/________________________________________> time

Pattern: repeated crossings of the set point. In physiology, this can appear as rhythmic fluctuations when the loop is “too aggressive” for its inherent delays.

Sensitivity and response speed: what changes the observed pattern

Sensor sensitivity

Sensitivity is how much the sensor signal changes for a given change in the controlled variable. If sensitivity increases, the controller “sees” a larger error signal for the same actual deviation, which effectively increases loop gain.

• Higher sensor sensitivity → larger perceived error → stronger corrective drive → faster correction but more overshoot/oscillation risk.
• Lower sensor sensitivity → smaller perceived error → weaker corrective drive → slower correction and larger steady deviations may persist.

Effector strength (actuator effectiveness)

Effector strength is how much the controlled variable changes per unit of command (e.g., how strongly a gland secretes in response to a signal, or how powerfully a muscle changes flow/pressure for a given neural input). Stronger effectors also raise effective gain.

• Stronger effector → bigger output change per command → faster correction but higher overshoot risk.
• Weaker effector → smaller output change per command → slower correction, may fail to restore near the set point under stress.

Response time and delays

Response time is how quickly the controlled variable moves back toward the set point after a disturbance. It depends on gain and on delays/inertia:

• High gain tends to shorten response time.
• Long delays (slow hormone onset, slow diffusion, slow mechanical changes) increase overshoot/oscillation risk because the system keeps “pushing” based on outdated information.

Trade-off: fast vs. accurate (and stable)

Many physiological loops face a trade-off:

• Fast correction reduces time spent away from the set point but can increase variability (overshoot, oscillations).
• Accurate/stable correction minimizes swings but may allow longer periods of deviation.

A useful mental model is to separate two goals: (1) speed (how quickly error shrinks), and (2) smoothness (how little the system overshoots). Raising gain often improves (1) while harming (2), especially when delays are present.

Practical step-by-step: predicting patterns from parameter changes

Use this three-step method whenever you are told “sensor sensitivity increased” or “effector became stronger.”

1. Translate the change into effective gain: does the loop respond more strongly to the same deviation? If yes, effective gain increased.
2. Check for delay/inertia: if the pathway is slow (transport, hormone kinetics, tissue lag), higher gain is more likely to overshoot or oscillate.
3. Predict the time-course shape: low gain → slow monotonic approach; moderate gain → quick approach with little/no overshoot; high gain (especially with delay) → overshoot and oscillations.

(3) Reasoning drills: predict the output pattern

Each drill describes a disturbance at time 0. Your task is to predict the qualitative graph shape (slow correction, overshoot, oscillation) and whether response time increases or decreases.

Drill 1: increased sensor sensitivity, same effector, same delays

• Change: sensor output per unit deviation increases.
• Predict: effective gain increases → faster initial correction; higher chance of overshoot. If delays are nontrivial, oscillation risk increases.
• Expected pattern: shift from Graph 1 toward Graph 2 or 3.

Drill 2: decreased sensor sensitivity, same effector, same delays

• Change: sensor under-reports deviations.
• Predict: effective gain decreases → slower correction; less overshoot; larger residual deviation may persist longer.
• Expected pattern: shift toward Graph 1 (slower, smoother).

Drill 3: stronger effector (more output change per command), same sensor

• Change: actuator is more powerful.
• Predict: effective gain increases → faster correction; overshoot more likely; oscillation more likely if delays exist.
• Expected pattern: Graph 2 or 3 depending on delay magnitude.

Drill 4: weaker effector, same sensor

• Change: actuator produces smaller changes for the same command.
• Predict: effective gain decreases → slower correction; overshoot less likely; may fail to return close to set point during ongoing disturbance.
• Expected pattern: Graph 1 with prolonged deviation.

Drill 5: same gain, increased delay (slower transport or slower tissue response)

• Change: information or effect arrives later.
• Predict: even without changing gain, the system behaves as if it is “too aggressive for its timing” → more overshoot and oscillation.
• Expected pattern: shift toward Graph 2 or 3 (more ringing).

Drill 6: sensor sensitivity increases but effector strength decreases (opposing changes)

• Change: sensor amplifies error signal, but actuator produces less change per command.
• Predict: net effect depends on which change dominates. If sensor increase outweighs effector decrease, effective gain rises (faster, more overshoot). If effector weakening dominates, effective gain falls (slower, smoother).
• How to decide: imagine a fixed deviation (e.g., error = 10). If the sensor/controller output doubles but the effector response halves, the net output change is similar → pattern may be similar but could become noisier if the sensor is more reactive.

Drill 7: effector becomes stronger and also slower (more powerful but with longer time constant)

• Change: bigger eventual correction, but it ramps up and down slowly.
• Predict: initial response may not be faster; overshoot risk can increase because the effector keeps changing after the error has already shrunk (inertia). This combination often produces delayed overshoot and oscillation.
• Expected pattern: slower start, then overshoot (a delayed version of Graph 2/3).

Quick reference table: parameter change → expected behavior