• forced desynchrony protocol;
  • frontal low electroencephalogram activity;
  • homeostasis;
  • plasma melatonin;
  • spectral analysis;
  • spindle activity


Electroencephalographic (EEG) activity is a key indicator of a vigilance state, and quantitative analyses of the EEG have revealed profound differences both between and within vigilance states in humans. We summarize recent studies that investigated how the spectral composition of the EEG during the three vigilance states, that is, wakefulness, rapid eye movement (REM) and non-REM sleep, is modulated by a circadian oscillator, which is independent of sleep–wake behavior, and by the sleep–wake oscillation itself, that is, elapsed time awake and elapsed time asleep. The data collected in sleep deprivation experiments and in protocols in which the sleep–wake cycle was desynchronized from endogenous circadian rhythmicity show that both factors contribute to this variation in a frequency- and state-specific manner. Low frequency EEG activity, including slow waves and theta frequencies, during both wakefulness and non-REM sleep, gradually increases with elapsed time awake and progressively declines with elapsed time asleep. The EEG activity in this 0.75–8 Hz frequency range is not markedly affected by circadian phase. In contrast, alpha activity (8–12 Hz) during wakefulness and REM sleep, as well as sleep spindle activity (12–15 Hz) during non-REM sleep, show a robust circadian regulation. Circadian and sleep–wake dependent regulation of EEG activity within the vigilance states also exhibits topographical variation such that frontal brain areas are more susceptible to the effects of the sleep homeostat than more parietal brain regions. It will be challenging to identify the functional correlates of these different spectral EEG patterns and relate them to neurobehavioral performance and recovery functions of sleep.