INFORMATION MEASURES QUANTIFYING APERIODIC STOCHASTIC RESONANCE

Aperiodic stochastic resonance (ASR) is a phenomenon in which the resp onse of a nonlinear system to a subthreshold information-bearing signa l is optimized by the presence of noise. We have previously characteri zed this effect by the use of cross-correlation-based measures. Here w e apply a measure (transinformation) that directly quantifies the rate or information transfer from stimulus to response and show that the p resence of noise optimizes the information-transfer rate. By consideri ng a nonlinear system (the FitzHugh-Nagumo model) that captures the fu nctional dynamics of neuronal firing, we demonstrate that sensory neur ons could, in principle, harness ASR to optimize the detection and tra nsmission of weak stimuli.