Healthy brains exhibit a rich dynamical repertoire with flexible and varied spatiotemporal patterns replays on both microscopic and large scales. Neurodegenerative diseases reduce this functional repertoire. We hypothesize that microscopic dynamics must operate in or near a critical regime for the functional repertoire to be fully explored and for realistic flexible dynamics to emerge. To test this hypothesis, we use a modular Spiking Neuronal Network model, where each group of Leaky Integrate and Fire neurons represents a cortical region. A STDP-based rule is used to learn patterns of activations which propagate between modules based on a probabilistic distribution linked to the quantity of white-matter fibers for long-range connections and an exponential decay rule for nearby regions. Consequently, information (i.e., the learned spatiotemporal modular patterns) is encoded both as a precise sequence of neurons and as a trajectory of activated modules. The model displays two distinct dynamical regimes: an uncorrelated low-rate state and a strongly correlated state, marked by a high Order Parameter value (indicating the similarity of spontaneous activity with stored patterns). These regimes are separated by either a first-order or second-order phase transition, depending on the strength of global inhibition and structured connections. When the hysteresis loop shrinks, a continuous phase transition occurs, and at the edge between the two states it opens up an extended region with high order parameter fluctuations. This region manifests as numerous transients and intermittently reoccurring motifs events. The model's predictions are compared with empirical data from MEG recordings in healthy adults. Notably, a high correlation between the functional dynamics (measured by the distribution of regions which initiate avalanches) in synthetic and MEG data is observed only when the model operates in the critical extended region.