PII S0361-9230(98)00018-5REVIEW ARTICLEVertebrates that never sleep: Implications for sleep’sbasic functionJ. Lee Kavanau*University of California, Department of Biology, Los Angeles, CA, USA[Received 13 October 1997; Revised 5 February 1998; Accepted 10 February 1998]ABSTRACT: A major activity of the brain of most vertebratesduring waking behavior is the processing of sensory informa-tion, preponderantly visual. This processing is not fully compat-ible with the brain’s spontaneous oscillatory activity that main-tains (refreshes) infrequently used circuits that store inheritedand experiential information (memories). Great reduction insensory input and processing during sleep permits the refresh-ment of memory circuits to occur unimpededly. Accordingly,sleep may have evolved as ever augmenting needs for process-ing visual information during waking behavior by brains of greatcomplexity conflicted increasingly with needs to refresh mem-ory circuits. The lack of a need for sleep by genetically blindfishes that live in caves, and sighted fishes that swim continu-ously, is consistent with this thesis, as their needs for process-ing of sensory information, predominantly visual, are eithergreatly reduced or nil. Reduced requirements for processingsensory information by continuously swimming fishes owe tothe following aspects of their behavior and ecology: (1) visualinput is greatly reduced or absent during lengthy periods ofnocturnal activity; (2) schooling greatly reduces needs for sen-sory information, particularly visual; (3) being maintainedthrough frequent use, circuitry for most inherited memoriesneeds no refreshment; and (4) inasmuch as they lead a com-paratively routine existence in essentially featureless, open wa-ters, pelagic species acquire, and have need to refresh, rela-tively few experiential memories. Analogous circumstancescould account for the ability of migrating birds to fly for dayswithout rest or sleep. © 1998 Elsevier Science Inc.KEY WORDS: Vision and sleep, Memory circuit refreshment,Schooling functions, Continuous swimming, Sharks, Reef-dwelling fishes, Scombrids, Troglobionts.INTRODUCTIONEfforts to identify sleep’s basic function are complicated by an-cillary benefits in mammals and birds, such as bodily rest andrejuvenation, physiological restoration, regulation of hormonalsecretions, and reinforcement of the immune system [29,140].Such efforts can be more sharply focused by identifying theecobehavioral property common to all nonsleeping vertebrates:they spend much or all of their lives under conditions of reduced,minimal or nil visual input and, for many of them, also withreduction of other sensory inputs.These findings raise the possibility that the selective pressurefor the evolutionary origin of sleep was a conflict between thebrain’s waking needs for processing sensory information, predom-inantly visual, with its needs to refresh memory circuits for infre-quently used functions. The conflict might have arisen as non-sleeping animals acquired increasingly complex brains, behavior,and visual competencies, together with ever enlarging stores ofexperiential and inherited memories. If under selective pressure,the brain were to achieve a more profound state of unresponsive-ness to sensory inputs than usually occurs during restful waking,namely the state of sleep, circuit refreshment could have proceededunimpededly. Evidence consistent with this thesis is presented inthis article.Nonsleeping vertebrates encompass a wide range. Almost all ofthem swim continuously. A few rest most of the time. Some aregenetically blind, others have excellent visual acuity, still othershave great visual sensitivity at low light levels but their acuity ispoor. They cover the gamut from being color blind to havingfour-pigment color vision and a greater diversity of visual pig-ments than all other vertebrates combined. Some can occlude theirpupils, others have only limited occular control over ambient lightreception.Adult weights range from several g to hundreds of kg. Somehunt by day, some by night, others take food at any time. Someinhabit shallow waters and coral reefs, others are pelagic, range farinto the great ocean basins, and migrate between the continents.Some are strict ectotherms (loosely speaking cold-blooded), othersare “partially warm-blooded”, maintaining their bodies at temper-atures above external ambient values.With the exception of some salamanders, all nonsleeping ver-tebrates considered here are fishes. Several studies indicate that afrog and three reptiles also do not sleep, engaging only in restfulwaking [12]. Inasmuch as they provide less clear-cut examples,they are not considered further. As employed here, “restful wak-ing” or “rest” is characterized minimally by behavioral quiescence,* Address for correspondence: J. Lee Kavanau, University of California, Department of Biology, 405 Hilgard Avenue, Los Angeles, CA 90095-1606,USA. Fax: (310) 206-3987; E-mail: [email protected] Research Bulletin, Vol. 46, No. 4, pp. 269–279, 1998Copyright © 1998 Elsevier Science Inc.Printed in the USA. All rights reserved0361-9230/98 $19.00 1 .00269characteristic postures, vigilance, reduced or absent complex vi-sual processing, and unaltered sensory thresholds. It usually isengaged in only under conditions of relative safety, in which thereis little need for close monitoring of sensory information.REFRESHMENT OF MEMORY CIRCUITS ANDPROCESSING OF EXOGENOUS INFORMATIONSpontaneous, endogenous activities of the brain during sleepare treated here, with a discussion of the basis for the conflictbetween these activities and exogenously-induced processing ofsensory input, chiefly complex visual information. Examples aredrawn largely from mammals [56–58].Spontaneous Brain Activities During SleepIn 1966, Roffwarg et al. [108] proposed that spontaneous,repetitive activations of circuitry in the central nervous system(CNS) of the human embryo during rapid-eye-movement (REM)sleep facilitate the development and maintenance of inherited(genetically programmed) memory circuits. They suggested thatsuch activations during REM sleep maintain (refresh) inheritedcircuitry throughout life. Subsequent investigators extended theconcept to include refreshing of circuits storing experiential infor-mation [46,100,124].Functional and “nonutilitarian” “dynamic stabilization.” Theconcept provided the basis for a paradigm of “dynamic stabiliza-tion” (DS) of