5.9 Feedback Threshold

A feedback threshold is a critical level or boundary of an error, deviation, or state variable that determines whether a feedback mechanism activates, changes its mode of operation, or produces a qualitatively different response. Rather than responding continuously and proportionally to all deviations, a system with a feedback threshold behaves differently below and above that threshold: small deviations may produce no response, a weak response, or a qualitatively different response compared to deviations that cross the threshold. Thresholds introduce nonlinearity into feedback systems and are responsible for many important biological, physical, and social phenomena including switching, hysteresis, and catastrophic transitions.

The simplest form of a feedback threshold is the dead zone or dead band, a range of error values around the setpoint within which no corrective action is taken. This is common in practical control systems to avoid unnecessary actuator activity in response to small errors within acceptable tolerance limits. When the error exceeds the dead band boundaries in either direction, the feedback loop activates and corrective action begins:

where e_th is the threshold error magnitude. This type of control is used in relay control systems, on-off thermostats, and bang-bang controllers, where the actuator has only two states (fully on or fully off) rather than continuous variable output.

Biological feedback thresholds are central to the functioning of sensory and signaling systems. The action potential in neurons is driven by a threshold in membrane voltage: small depolarizations below the threshold fail to trigger an action potential and decay passively, while depolarizations that reach or exceed the threshold trigger the rapid, all-or-nothing spike. This threshold behavior results from the cooperative opening of voltage-gated sodium channels when the membrane potential crosses approximately −55 mV, creating a positive feedback loop that drives the potential to near the sodium equilibrium potential. The threshold ensures that only sufficiently strong signals propagate through the nervous system, filtering out small noise fluctuations.

The threshold in the immune system between self-tolerance and immune activation is another critical biological feedback threshold. T cells are activated only when the signal from their antigen receptor, combined with co-stimulatory signals, exceeds a threshold level. Below this threshold, the response is suppressed to prevent autoimmune reactions. Above the threshold, the adaptive immune response is triggered. This thresholded feedback prevents the immune system from attacking self-tissues while remaining sensitive to genuine pathogens.

Tipping points in complex systems are large-scale expressions of feedback thresholds. A tipping point is a threshold value of a state variable or parameter beyond which the system undergoes a qualitative change in behavior, often driven by a shift from stabilizing to destabilizing feedback. Examples include the threshold atmospheric CO₂ concentration beyond which the Greenland ice sheet becomes committed to melting, the social contagion threshold beyond which an opinion or behavior spreads through a population without further external forcing, and the financial leverage threshold beyond which debt-deflation dynamics create a self-reinforcing collapse.

Hysteresis in thresholded feedback systems means that the threshold for activation and the threshold for deactivation may be different, so that the system's behavior depends on its history as well as its current state. A thermostat with hysteresis turns the heater on when temperature drops below T_low and turns it off when temperature rises above T_high, with a dead band between T_low and T_high. Once the heater is on, it stays on until T_high is reached; once it is off, it stays off until T_low is reached. This hysteresis prevents rapid switching that would wear out the actuator, but means the system tolerates a wider range of temperature variation than a zero-hysteresis controller.

Setting feedback thresholds appropriately is a design challenge in all systems where they appear. Thresholds set too high cause the system to ignore significant deviations that should be corrected, allowing large errors to accumulate. Thresholds set too low cause unnecessary actuation in response to noise, wasting energy, wearing out actuators, and potentially introducing oscillations. In biological and social systems, threshold calibration is an ongoing adaptive process driven by experience, learning, and selective pressure, rather than a one-time engineering specification.