We first study here a simple stochastic dynamical equation with delayed self-feedback. The model is investigated numerically and we find that its dynamics show an emergent regular “spiking” behavior, by “tuning” its “noise” and “delay”. In order to gain insight into this “resonance”, we abstract the model and study a stochastic binary element whose transition probability depends on its state at a fixed interval in the past. With this abstracted model we can analytically capture the time interval histograms between spikes, and discover how the resonance between noise and delay arises. The resonance is also observed when such stochastic binary elements are coupled through delayed interaction. ©2000 American Institute of Physics.