Noise actually helps detect a weak periodic signal in a bistable potential. Optimal noise level maximizes SNR — the defining feature of stochastic resonance.
Stochastic resonance: In a bistable/threshold system, adding the right amount of noise enhances detection of a subthreshold periodic signal.
Kramers escape rate: r_K ~ exp(-ΔU/D)
Optimal noise matches half-period: T/2 ~ 1/r_K
SNR peak at D* ~ ΔU/ln(signal period)
Applications: neuroscience (hair cells, neurons), climate (ice ages), signal processing. Discovered by Benzi et al. (1981) for ice-age periodicity.