Whereas usually deemed to be mediated by synaptic plasticity, the formation of long-term memory may rely on dormant immature neurons distributed over the cortex, especially in mammals with large brain. Notably owing to axonal sprouting, these precursors of cortical cells could be engaged with subcortical neurogenesis into the brain lifelong adaptation to changing environment. On the computational modelling side, Guided Propagation memory models implement coincidence-detectors that are dynamically recruited for the experience-driven growth of long-lasting memory paths. The latter are meant to integrate/generate trajectories of discrete events at the basis of sequential learning and behaviour, known as a cortical function (Cx). To account as well for short-term memory, renewable phase-coding is supported by generic paths whereby serial patterns can instantly be retained and replayed (likely a hippocampal function: Hp). When repeatedly broadcasted through inner modulation signals during a suitable anticipation-free state, namely offline, every such trajectory recorded by Hp may give rise-via the Thalamus-to permanent traces encoded within Cx channels, provided unused coincidence-detectors at disposal. Conversely, cross-links built between parallel channels notably guide the learning of embedded trajectories, and allow Cx to inform Hp about their novelty in order for Hp to record unexpected events. Potential neural contributors to learning are addressed here from this collaborative perspective, including hippocampal and cortical replay, electric oscillations and transients, monoaminergic and ionic signalling. Computer experiments show how phasic/tonic and activation/modulation signals can together support the translation of Hp temporary characteristic phases into Cx memory maps over series of simulated online/offline cycles. By projecting the causal relationships linking the model components onto neurobiological data, predictions are made about processes which underlie learning and memory. The issue of whether quiescent cortical cells in reserve could be embedded in the working cortex in due course is eventually investigated in the light of recent molecular and anatomical insights, which elicits a functional matching between the proposed architecture and Hippo-thalamo-cortical circuits. There thus emerges a form of lifelong brain development wherein thalamic waves accompanying the hippocampus offline reactivation of previous online experience coincide with cortical spontaneous activity so as to grow cortical maps which combine existing and free canonical circuits.