Roles of cortical neuromodulation and offline reactivation in memory consolidation of emotionally salient visual experiences

Summary What are the mechanisms by which we selectively remember sensory cues associated with salient outcomes? Salient sensory experiences activate distributed patterns of neurons in regions including sensory cortex, amygdala, and hippocampus. The linking of neural representations of cues and outcomes is believed to occur both online, during the sensory experiences, and offline, during joint reactivation of cue- and outcome-related patterns of neurons during subsequent quiet periods. Both the online and offline aspects of cue-outcome association learning have been posited to involve co-activation of neuromodulators that project to these distributed brain regions, which may facilitate neural plasticity via actions on intracellular signaling in neuronal dendrites. We have recently developed methods to visualize these network, cellular and subcellular processes across days and weeks in lateral visual association cortex (LVAC) of behaving mice learning an operant Go-NoGo visual discrimination task. LVAC is a key hub that links the hippocampus, sensory cortex, and amygdala. Silencing of either online or offline activity in LVAC perturbs long-term memory consolidation and recall of cue-outcome associations. We previously found that LVAC neuron cue responses are highly plastic across learning and that LVAC is necessary for performance of our task. We showed that the same distinct pattern of neurons that was activated by a given visual cue was subsequently reactivated for ~100-200 ms during quiet waking, with higher reactivation rates during early learning and for salient food- predicting vs. neutral cues. Rates of food-cue but not neutral cue reactivation predicted next-day improvements in performance. Accordingly, reactivations involving ensembles of neurons encoding both the food cue and reward predicted strengthening of next-day functional connectivity of participating neurons, providing a potential cortical substrate for associative learning. This proposal seeks to define the neuromodulatory, signaling, and network mechanisms underlying the above findings, and their causal role in learning. A brain region that densely innervated LVAC and implicated in salience, plasticity, and memory consolidation is the locus coeruleus (LC). Both online and offline LC activity guides which salient experiences are stored. Noradrenaline released by LC neurons can act on beta-adrenergic receptors, thereby boosting cyclic AMP (cAMP), a second messenger that facilitates synaptic plasticity. Thus, we hypothesize that increased neuromodulatory input from LC to LVAC during salient sensory experiences (Aim 1) drives increases in cAMP signaling in LVAC neurons (Aim 2), and biases subsequent offline reactivation of neurons activated during these experiences, in order to modify next-day changes in network activity and behavioral performance (Aims 1-2). We then examine electrophysiological correlates of offline cortical reactivations and their causal role in facilitating network changes and learning (Aim 3). These cellular/subcellular investigations of neural activity and signaling in vivo will address links between neuromodulation, offline reactivation and associative plasticity.