Recent studies suggest that sleep differentially alters the activity of cortical neurons based on firing rates during preceding wake—increasing the firing rates of sparsely firing neurons and decreasing those of faster firing neurons. Because sparsely firing cortical neurons may play a specialized role in sensory processing, sleep could facilitate sensory function via selective actions on sparsely firing neurons. To test this hypothesis, we analyzed longitudinal electrophysiological recordings of primary visual cortex (V1) neurons across a novel visual experience which induces V1 plasticity (or a control experience which does not), and a period of subsequent ad lib sleep or partial sleep deprivation. We find that across a day of ad lib sleep, spontaneous and visually-evoked firing rates are selectively augmented in sparsely firing V1 neurons. These sparsely firing neurons are more highly visually responsive, and show greater orientation selectivity than their high firing rate neighbors. They also tend to be “soloists” instead of “choristers”—showing relatively weak coupling of firing to V1 population activity. These population-specific changes in firing rate are blocked by sleep disruption either early or late in the day, and appear to be brought about by increases in neuronal firing rates across bouts of rapid eye movement (REM) sleep. Following a patterned visual experience that induces orientation-selective response potentiation (OSRP) in V1, sparsely firing and weakly population-coupled neurons show the highest level of sleep-dependent response plasticity. Across a day of ad lib sleep, population coupling strength increases selectively for sparsely firing neurons—this effect is also disrupted by sleep deprivation. Together, these data suggest that sleep may optimize sensory function by augmenting the functional connectivity and firing rate of highly responsive and stimulus-selective cortical neurons, while simultaneously reducing noise in the network by decreasing the activity of less selective, faster-firing neurons.

Learning-activated engram neurons play a critical role in memory recall. An untested hypothesis is that these same neurons play an instructive role in offline memory consolidation. Here we show that a visually-cued fear memory is consolidated during post-conditioning sleep in mice. We then use TRAP (targeted recombination in active populations) to genetically label or optogenetically manipulate primary visual cortex (V1) neurons responsive to the visual cue. Following fear conditioning, mice respond to activation of this visual engram population in a manner similar to visual presentation of fear cues. Cue-responsive neurons are selectively reactivated in V1 during post-conditioning sleep. Mimicking visual engram reactivation optogenetically leads to increased representation of the visual cue in V1. Optogenetic inhibition of the engram population during post-conditioning sleep disrupts consolidation of fear memory. We conclude that selective sleep-associated reactivation of learning-activated sensory populations serves as a necessary instructive mechanism for memory consolidation.

36Learning-activated engram neurons play a critical role in memory recall. An untested 37 hypothesis is that these same neurons play an instructive role in offline memory consolidation. 38 Here we show that a visually-cued fear memory is consolidated during post-conditioning sleep 39 in mice. We then use TRAP (targeted recombination in active populations) to genetically label or 40 optogenetically manipulate primary visual cortex (V1) neurons responsive to the visual cue. 41Following fear conditioning, mice respond to activation of this visual engram population in a 42 manner similar to visual presentation of fear cues. Cue-responsive neurons are selectively 43 reactivated in V1 during post-conditioning sleep. Mimicking visual engram reactivation 44 optogenetically leads to increased representation of the visual cue in V1. Optogenetic inhibition 45 of the engram population during post-conditioning sleep disrupts consolidation of fear memory. 46 We conclude that selective sleep-associated reactivation of learning-activated sensory 47 populations serves as a necessary instructive mechanism for memory consolidation. 48 49Introduction 50 51Experiences during wake influence neural activity patterns during sleep. For example, 52hippocampal place cells activated during environmental exploration in wake show higher firing 53 rates (reactivation) 1 and/or similar sequences of activity (replay) 2-6 during subsequent sleep. 54This phenomenon has been observed in multiple brain regions, multiple species, and following a 55 wide range of experiences 7-13 . Since sleep loss has a disruptive effect on many forms of 56 memory 14 , replay and reactivation may play an instructive role in sleep-dependent memory 57 consolidation 14,15 . To test this, prior work has disrupted network-wide activity during specific 58 sleep oscillations [16][17][18][19] or disruption of activity in genetically-defined cell types across specific 59 phases of sleep 20-23 -but not the specific neurons activated during learning itself. Recent work 60 has emphasized the essential role of engram neurons in memory recall 24,25 . To date however, 61 no studies have applied this technology to the question of sleep-dependent memory 62consolidation. 63Here we test the necessity of sleep-specific engram neuron reactivation for memory 64 consolidation. We describe a form of visually-cued fear memory in mice, which is encoded by 65 single trial conditioning (pairing presentation of an oriented grating visual stimulus with an 66 aversive foot shock) and dependent on post-conditioning sleep. Post-conditioning, the mice 67 behaviorally discrimination between conditioned and neutral visual cues, leading to a selective 68 fear memory. This discrimination is disrupted by post-conditioning sleep deprivation. Using this 69 paradigm, we take advantage of recently developed genetic tools to selectively manipulate 70 orientation-selective (i.e., cue-activated) primary visual cortex (V1) neurons. We find that these 71 cue-activated visual engram neurons are selectively rea...