Breakdown of stochastic resonance in complex networks

In networked systems, stochastic resonance occurs as a collective phenomenon where the entire stochastic network resonates with a weak applied periodic signal. Beyond the interplay among the network coupling, the amplitude of the external periodic signal, and the intensity of stochastic fluctuations, the maintenance of stochastic resonance also crucially depends on the resonance capacity of each oscillator composing the network. This scenario raises the question: Can local defects in the ability of oscillators to resonate break down the stochastic resonance phenomenon in the entire network? Here, we investigate this possibility in complex networks of prototypical bistable oscillators in a double-well potential. We test the sustainability of stochastic resonance by considering a fraction of network oscillators with nonresonant potential landscapes. We find that the number of nonresonant oscillators depends nonlinearly on their dissimilarity from the rest of the network oscillators. In addition, we unravel the role of the network topology and coupling strength in maintaining, or suppressing, the stochastic resonance for different noise levels and number of nonresonant oscillators. Finally, we obtain a low-dimensional deterministic model confirming the results observed for the networks.