Parameterizing the relationship between motor cortical reactivation during sleep and motor skill acquisition in the freely behaving marmoset

点击翻译按钮获取中文摘要

Project Summary/Abstract This project will provide a more nuanced and mechanistic model of the role of sleep in memory consolidation, particularly as it pertains to procedural motor skill acquisition in a non-human primate model. Motor skill learning delineated by enhanced speed, automaticity, and accuracy of a correlate strongly with the duration of non-REM (NREM) sleep. Neural reactivations of daytime neural activity preferentially occur during NREM, and disruptions in NREM sleep negatively impacts memory consolidation. Since neural reactivations are not perfect copies of daytime activity it is unclear what specific information about behavior and skill acquisition is being reactivated during sleep. Do reactivations reflect certain parts or kinematic variables of the motor behavior conducted during the day? Do changes in these reactivations predict certain features of future motor skill improvements? We will develop a model that parameterizes the relationship between reactivation and memory by measuring the dependence of motor skill learning on the number of reactivations, the fidelity of reactivations, and, most importantly, the decodability of these reactivations each night and over subsequent nights. That is, we will build decoding algorithms that accurately predict upper limb movements from neural activity during the day and then use these algorithms to identify if spiking that is specific to certain kinematic variables are preferentially reactivated. We will use the natural process of retrograde interference when a subject learns a second motor skill following the first skill at various inter-task intervals to manipulate reactivation and skill acquisition to more causally link reactivation to motor skill acquisition. Finally, our model will enhance the standard sleep-consolidation framework using network science based tools to identify circuit level changes: with a particular emphasis on higher order relationships between superficial and deep neurons that are predictive of motor skill learning. To do so we will use wireless neural recordings from motor cortex (M1) in unrestrained marmoset monkeys (Callithrix jacchus) will examine motor skill acquisition and sleep- induced memory consolidation of these skills. Multi-electrode arrays with multiple contacts in depth will allow us to systematically parameterize the interdependence of reactivations and network changes across cortical lamina in M1 with motor skill performance. In Aim 1, we will measure changes in M1 population dynamics across cortical lamina as monkeys engage in naturalistic and artificial motor skill acquisition tasks. In Aim 2, we will characterize reactivations of skill-related neuronal activity patterns in M1 during sleep with a focus on the behaviorally-relevant information content of these reactivations using population decoding methods and functional network techniques. Finally, in Aim 3, we will examine retrograde interference and sleep reactivation to naturally manipulate reactivation and skill acquisition. These aims will provide one of the first and most comprehensive examinations of the role of sleep-induced reactivations of behaviorally relevant multineuronal activity patterns in motor skill acquisition of the primate.