In healthy conditions, the brain’s activity reveals a series of intermingled oscillations, generated by large ensembles of neurons, which provide a func-tional substrate for information processing. How single neuron properties influence neuronal population dynamics is an unsolved question, whose solution could help in the understanding of the emergent collective behaviors arising during cognitive processes Here we consider a neural massmodel, which reproduces exactly the macroscopic activity of a network of spiking neurons. This mean-field model is employed to shade some lighton an important and ubiquitous neural mechanism underlying information processing in the brain: theθ-γcross-frequency coupling.In particular, we will explore in detail the conditions under which two coupled inhibitory neural populations can generate these functionally relevant coupled rhythms.