A module of Kuramoto oscillators — dozens or hundreds of coupled phase oscillators within a densely connected subnetwork — can be replaced by a single effective oscillator. The condition: intra-module synchronization must be strong relative to inter-module coupling.
When this condition holds, the coarse-grained network of effective nodes accurately predicts phase transitions between synchronized and unsynchronized states. The simplified system preserves the global dynamics — the transitions happen at the same critical coupling strengths, for the same structural reasons, with the same qualitative behavior.
Bosnardo and de Aguiar demonstrate this for networks with 2-3 modules, showing that the prediction is independent of how the modules are interconnected. The topology between modules does not need to be known precisely — only the existence and relative strength of inter-module coupling matters.
The method is motivated by neuroscience, where EEG electrodes average over millions of neurons. Each electrode reading is already a coarse-grained effective oscillator. The formal result validates the approximation: when internal synchrony is strong, the averaging that electrodes perform does not distort the dynamical structure of the system. The measurement artifact is the correct theoretical object.
The structural insight generalizes: any hierarchically organized oscillating system where local synchronization is faster than global coupling admits a separation of scales. The fast internal dynamics average out, leaving effective degrees of freedom that interact on slower timescales. Coarse-graining is not information loss — it is the identification of the relevant variables.