An informative view of suprathreshold stochastic resonance with stimulus-specific information

The Shannon mutual information between the input stimuli and array responses is explicitly a weighted average information gain over individual contributions from the stimulus ensembles to the responses, but it does not address which input stimuli are more effectively encoded. In this paper, an informative measure of the stimulus-specific information (SSI) will be used to decompose the mutual information, and provide a microscopic insight into the system encoding. The SSI histograms not only determine the most efficient stimulus, but also provide an informative view of the noise benefit in information gains via the suprathreshold stochastic resonance mechanism.