Observations on delta homeostasis, the one-stimulus model of NREM-REM alternation and the neurobiologic implications of experimental dream studies.

We first review the concept of delta homeostasis as formulated in our 1974 model and as developed quantitatively by Borbely and colleagues in several versions of the two-process model. We illustrate difficulties in the application of this quantitative model to the negative delta rebound in the rat and we put forward additional evidence that the negative rebound is a pathological rather than a homeostatic response to sleep deprivation. We next review experiments on rats in which the waking metabolic rate of limbic structures was increased by blockade of the NMDA-gated cation channel with ketamine and MK-801. As predicted by the 1974 homeostatic model, NREM delta increased during subsequent sleep. However, it remains to be shown that this powerful effect is actually caused by the metabolic change and that it is an intensification of physiological sleep rather than a non-specific increase in EEG slow waves caused by neurotoxicity. We then outline our one-stimulus model of NREM/REM alternation. In this model NREM sleep is induced by periodic (neuroendocrine?) pulses. These pulses increase delta EEG amplitude and density, depress arousal level and inhibit neural activity. When the strength of the pulsatile stimulus falls below a critical level, REM emerges as neuronal escape. Last, we discuss the neurobiologic implications of two robust findings in experimental dream studies: the relation of dream reports to arousal and the consistent failure of controlled studies to demonstrate qualitative differences between NREM and REM mentation.