Fig. 9

We illustrate the emergence of SSR in the STS model for the same parameters as in Fig. 6d. a A fixed realization of the signal (blue) and three realizations (different common noise realizations) of the output of the STS population for and (black, green, and red). For better visualization the output is convoluted with a Gaussian filter and all outputs and the signal are rescaled to unit variance and zero mean over the time window shown. For vanishing independent noise, the individual spike trains of the population are identical for a fixed realization of the signal and the common noise. In this case signal transmission is not improved by the large population size. b Same as in a but for (close to the point of stochastic resonance). The individual noise leads to shifting of spikes, such that the convoluted summed output is smoothed. Note that the three realizations of the output are all close to the input signal as well as to each other, indicating a reliable signal transmission. c Same as in a but for (far beyond the point of stochastic resonance). Note that our average over a comparatively short time window implies the suppression of long-term variability (corresponding to leaving out low-frequency components of the coherence function)