Reservoir computing allows recovering hidden network dynamics

In this study, we examined reservoir computing (RC)as a tool for predicting the macroscopic dynamics of a subset of oscillators in a network based on the macroscopic dynamics of other parts of it. As a model network, we utilized a network of 300 Kuramoto oscillators with adaptation. Our results demonstrate that reservoir computing effectively addresses this task. Additionally, a similar reservoir computing model was applied to experimental neurovisualization data and exhibited high accuracy in reconstructing damaged EEG channels compared to classical methods like spatial interpolation.