U.S. flag

An official website of the United States government

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Purves D, Augustine GJ, Fitzpatrick D, et al., editors. Neuroscience. 2nd edition. Sunderland (MA): Sinauer Associates; 2001.

  • By agreement with the publisher, this book is accessible by the search feature, but cannot be browsed.
Cover of Neuroscience

Neuroscience. 2nd edition.

Show details

The Possible Functions of REM Sleep and Dreaming

Despite this wealth of descriptive information about the stages of sleep, the functional purposes of the various sleep states are not known. Whereas most sleep researchers accept the idea that the purpose of non-REM sleep is at least in part restorative, the function of REM sleep remains a matter of considerable controversy.

A possible clue about the purposes of REM sleep is the prevalence of dreams during these epochs of the sleep cycle. The occurrence of dreams can be tested by waking volunteers during either non-REM or REM sleep and asking them if they were dreaming. Subjects awakened from REM sleep recall elaborate, vivid, hallucinogenic and emotional dreams, whereas subjects awakened during non-REM sleep report fewer dreams, which, when they occur, are more conceptual, less vivid and less emotion-laden.

Dreams have been studied in a variety of ways, perhaps most notably within the psychoanalytic framework of revealing unconscious thought processes considered to be at the root of neuroses. Sigmund Freud's The Interpretation of Dreams, published in 1900, speaks eloquently to the complex relationship between conscious and unconscious mentation. It is by no means agreed upon, however, that dreams have the deep significance that Freud and others have given them, and the psychoanalytic interpretation of dreams has recently fallen into disfavor. Nevertheless, most people probably give some credence to the significance of dream content, at least privately. In more recent studies of dreams, about 65% are associated with sadness, apprehension, or anger; 20% with happiness or excitement; and, somewhat surprisingly, only 1% with sexual feelings or acts.

Adding to the uncertainty about the purposes of REM sleep and dreaming is the fact that deprivation of REM sleep in humans for as much as two weeks has little or no obvious effect on behavior. Such studies have been done by waking volunteers whenever their EEG recordings showed the characteristic signs of REM sleep. Although the subjects in these experiments compensate for the lack of REM sleep by having more of it after the period of deprivation has ended, they suffer no obvious adverse effects. Similarly, patients taking certain antidepressants (MAO inhibitors) have little or no REM sleep, yet show no obvious ill effects, even after months or years of treatment. The apparent innocuousness of REM sleep deprivation contrasts markedly with the effects of total sleep deprivation (see earlier). The implication of these several findings is that we can get along without REM sleep, but need non-REM sleep in order to survive.

Several general hypotheses about dreams and REM sleep have been advanced. Francis Crick (of DNA fame) and Grahame Mitchison suggested that dreams act as an “unlearning” mechanism, whereby certain modes of neural activity are erased by random activation of cortical connections. The hypothesis is based on the idea that the human brain represents information by the activity of sets of neuronal networks that are widely distributed and overlapping. In computers, neural network architectures are subject to unwanted patterns of activity that can indeed degrade rather than enhance the information content of the system. By analogy, these “parasitic” modes of activity might be unwanted thoughts or erroneous information, which, if not expunged, could become the basis for obsession, paranoia, or other pathologies of thought that prevent the “system” from working as efficiently as it should. In a different vein, Michel Jouvet proposed that dreaming reinforces behaviors not commonly encountered during the awake state (aggression, fearful situations) by rehearsing them while dreaming. Yet another hypothesis is that REM sleep and dreams are involved in the transfer of memories between the hippocampus and neocortex. Finally, it has been suggested that dreaming is simply an incidental consequence of REM sleep. None of these ideas are generally accepted.

In short, the questions of why we have REM sleep and why we dream remain unanswered, as are questions about consciousness per se (Box D).

Box Icon

Box D


By agreement with the publisher, this book is accessible by the search feature, but cannot be browsed.

Copyright © 2001, Sinauer Associates, Inc.
Bookshelf ID: NBK11121


  • Cite this Page
  • Disable Glossary Links

Related information

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...