How do billions of neurons cooperate to process information and mediate cognitive function such as learning, planning or memory? An influential hypothesis is that groups of neurons can synchronize within brief time windows to represent a cognitive entity or a concept (e.g. an apple, a sonata, the thought of my grand-mother; Hebb 1949). These groups, called `cell assemblies’ are thought to be the functional units of the brain. Recent advances in large-scale neuronal recordings enabled to investigate such synchronous activity within single areas of the brain. Yet, the functional relevance of these ensembles remains largely unexplored, especially in higher order brain areas. To address this question, we designed two research projects based on extracellular recordings in rats. First, we demonstrated and characterized synchronous activity spanning different parts of the prefrontal cortex and the striatum, constituting the cortico-basal ganglia loop. These distributed cell assemblies exhibited behaviorally correlated activation during a set-shifting task, and synchrony emerged when assembly members shifted their phase relative to ongoing brain rhythms. We propose that cross-structural assemblies are likely to be a general mechanism that can span across multiple brain areas and networks, therefore underlying highly integrated representations. Second, we analyzed the neural activity from hundreds of single units in the prefrontal-amygdalar networks during sleep to investigate the functional relevance of cell assemblies from the perspective of a downstream neuron. Our results suggest that the synchronous activity of cell assemblies facilitate and amplify the discharge of downstream neurons, termed ‘readers’ of the assembly. Interestingly, we showed that the assembly-reader communication changes with learning.