Stochastic resonance

Stochastic resonance refers to a phenomenon that can occur in a noisy nonlinear system with certain characteristics when it is excited by a periodic signal.

The term was coined in 1981 by Italian and Belgian physicists to explain the periodic recurrence of ice ages.

Stochastic resonance is of technical importance in the amplification and detection of periodic signals, which are very weak compared to (usually annoying ) system noise.

Similar to the phenomenon of resonance, where there is a best excitation frequency, it is in the stochastic resonance, an intensity of the noise in the signal can best be detected. This intensity is paradoxically not zero.

Example

The non-linear system in the neuronal motivated example shown in the illustration is the threshold (blue), above which an action potential is initiated. The signal ( red) always extends below the threshold. But if noise added to the signal (black), so the threshold is exceeded occasionally. Exceeding the threshold are represented by the blue lines at the bottom. The probability is higher than in the minima in the maxima of the noise-free signal. At very low noise intensity, the threshold is never exceeded at very high, however, the signal is no longer good in the dense sequence of action potentials from. A mean noise intensity is optimal for the image of the signal after the non-linear stage.