The Fine Structure of Slow-Wave Sleep Oscillations: from Single Neurons to Large Networks

The discovery that the electrical activity of the brain oscillates during sleep is almost as old as the discovery of the electroencephalogram (EEG). The first human EEG recordings already reported a propensity to show oscillations, of which type, frequency and amplitude highly depend on behavioral state (Berger 1929; see Fig. 4.1). In an alert, awake subject, the EEG is dominated by low-amplitude fast activity (“desynchronized EEG”) with high-frequency oscillations (beta, gamma), whereas during slow-wave sleep, the EEG shifts to large-amplitude, slow oscillations. The early stage of slow-wave sleep is associated with the appearance of spindle waves, which occur at a frequency of 7 to 14 Hz. As sleep deepens, EEG waves with slower frequencies (0.1 to 4 Hz), including delta waves and slow oscillations, appear and progressively dominate the EEG. During paradoxical sleep, also called rapid-eye movement (REM) sleep, EEG activities are desynchronized and resemble those of wakefulness. Finally, some pathological states also display clear-cut oscillations, such as the “spike-and-wave” patterns (∼3 Hz) characteristic of many types of generalized epileptic seizures. The cellular bases of slow-wave sleep oscillations have been investigated since the first extracellular and intracellular recordings in mammals. The major brain regions which have been identified are the thalamus and cerebral cortex, which are intimately linked by means of reciprocal projections. The activities of thalamic and cortical neurons during sleep have been largely documented by electrophysiological studies. The cellular mechanisms underlying these oscillations depend on many