Cognitive neuroscience of sleep

Abstract

Mechanism is at the heart of understanding, and this chapter addresses underlying brain mechanisms and pathways of cognition and the impact of sleep on these processes, especially those serving learning and memory. This chapter reviews the current understanding of the relationship between sleep/waking states and cognition from the perspective afforded by basic neurophysiological investigations. The extensive overlap between sleep mechanisms and the neurophysiology of learning and memory processes provide a foundation for theories of a functional link between the sleep and learning systems. Each of the sleep states, with its attendant alterations in neurophysiology, is associated with facilitation of important functional learning and memory processes. For rapid eye movement (REM) sleep, salient features such as PGO waves, theta synchrony, increased acetylcholine, reduced levels of monoamines and, within the neuron, increased transcription of plasticity-related genes, cumulatively allow for freely occurring bidirectional plasticity, long-term potentiation (LTP) and its reversal, depotentiation. Thus, REM sleep provides a novel neural environment in which the synaptic remodelling essential to learning and cognition can occur, at least within the hippocampal complex. During non-REM sleep Stage 2 spindles, the cessation and subsequent strong bursting of noradrenergic cells and coincident reactivation of hippocampal and cortical targets would also increase synaptic plasticity, allowing targeted bidirectional plasticity in the neocortex as well. In delta non-REM sleep, orderly neuronal reactivation events in phase with slow wave delta activity, together with high protein synthesis levels, would facilitate the events that convert early LTP to long-lasting LTP. Conversely, delta sleep does not activate immediate early genes associated with de novo LTP. This non-REM sleep-unique genetic environment combined with low acetylcholine levels may serve to reduce the strength of cortical circuits that activate in the ~50% of delta-coincident reactivation events that do not appear in their waking firing sequence. The chapter reviews the results of manipulation studies, typically total sleep or REM sleep deprivation, that serve to underscore the functional significance of the phenomenological associations. Finally, the implications of sleep neurophysiology for learning and memory will be considered from a larger perspective in which the association of specific sleep states with both potentiation or depotentiation is integrated into mechanistic models of cognition.

Fig. 1

Electroencephalographic (EEG) and electromyographic (EMG) signals across 20 s of recording time. (A) Vertical grid spacing is 200 mV. Cortical EEG was taken as the differential signal from screw electrodes placed over the frontal cortices. Hippocampal EEG (hEEG) was recorded from a tetrode (4 individual 12 μm nichrome recording wires twisted together such that the distance between wires is ~15 μm) placed at the hilus and referenced to a tetrode placed in the corpus callosum. These data were taken from a rat implanted with a tetrode assembly (12 tetrodes) recorded from a resting pot after a maze run during a period of wakefulness (high variable EMG) without voluntary movement, allowing slow waves rather than theta to appear in the hEEG trace. Slow waves appear in the hippocampus both during non-REM sleep and during quiet wakefulness. (B) Times of action potentials from 17 simultaneously recorded hippocampal neurons (1 spike raster row for each cell) show bursts of coincident reactivation at the depolarized peaks of the hippocampal slow waves (arrows) during this period of quiet wakefulness.