Understanding the neurobiology of memory is recognised as one of the grand challenges of contemporary neuroscience.  It is approached from multiple levels of analysis – from ‘top-down’  (in recognising the several qualitatively different types of memory and their brain imaging correlates), through to ‘bottom-up’ approaches (such as both conserved and newly evolved molecular mechanisms of neuronal plasticity).  The two connected strands of this talk include, first, studies of the ‘making’ of memory (encoding) which is widely thought to involve unsupervised activity-dependent synaptic plasticity in memory trace formation, with memory allocation guided in part by differential spatial patterns of neuronal excitability.  New work from my own group has, however, called into question whether postsynaptic cell firing is always required for certain forms of memory encoding (Rossato et al, Current Biology, 2018).  Then, second, I will turn to the ‘keeping’ of memory (consolidation) with a focus on one specific aspect of memory retention.  This derives from the ‘synaptic tagging and capture’ model of protein synthesis-dependent long-term potentiation (Frey and Morris, Nature 1997) which asserts that, as much memory encoding is automatic, there must be selectivity to its retention to avoid saturation of the distributed networks of the hippocampus and related structures that mediate recent memory storage.  This leads on to identifying diverse determinants of selectivity, including peri-event novelty as one factor.  This can enhancing retention through both an immediate effect on attention and more generally via the enhancement of the synthesis of plasticity-related proteins.  This issue has recently been explored in mice using optogenetic techniques focusing, somewhat surprisingly, on the locus coeruleus (Takeuchi et al, Nature, 2016), a brain structure that has been the focus of important recent research in Tuebingen (Totah et al, Neuron 2018).