Sleepiness: evaluating and quantifying methods

Ž .International Journal of Psychophysiology 41 2001 251�263

Review

Sleepiness: evaluating and quantifying methods

Giuseppe Curcio�, Maria Casagrande, Mario BertiniDipartimento di Psicologia, Uni�ersita degli Studi di Roma ‘La Sapienza’, Via dei Marsi, 78, I-00185 Roma, Italy`

Received 19 June 2000; accepted 22 January 2001

Abstract

The aim of this literature review is to analyze the methods mainly used for evaluating and quantifying the complexphenomenon of sleepiness. The most common distinction is between subjective measures or self-evaluations,performance decrement measures, measures for evaluating sleep propensity and measures of arousal decrease.Techniques mainly used in specialized literature will be briefly presented and commented upon, evaluating their

Ž .sensitivity, advantages and limitations. We conclude that: a different measures inevitably are differently sensitive toŽ . Ž .sleepiness fluctuations; b the amount of prior sleep is strongly relevant in quantifying sleepiness levels; c

subjective and behavioral measures show a higher level of vulnerability to external and motivational factors. � 2001Elsevier Science B.V. All rights reserved.

Keywords: Sleepiness; Methods; Performance; Vigilance test; Self ratings; Physiological measures

1. Introduction

Sleepiness is unanimously considered to be oneof the main problems of modern society. Extremework rhythms, shift work and hostile environ-ments seem to favour its development, inducing aprogressive worsening of health status. Effectsare easily visible in several contexts: at work˚Ž . ŽAkerstedt, 1995 ; in the military Porcu et al.,`

. Ž1998 ; in road safety and public health Dinges,

� Corresponding author. Tel.: �39-06-49917508; fax: �39-06-4451667.

Ž .E-mail address: [email protected] G. Curcio .

.1995; Horne and Reyner, 1995 ; and daily andŽeducation environments Dement and Gelb,

.1993 .The concept of sleepiness is used by different

people to mean different things: languor or inert-Ž .ness Kleitman, 1963 ; subjective state of sleepŽ .need Broughton, 1989 ; a physiological drive re-

Ž .sulting from sleep deprivation Aldrich, 1994 ; orŽa strong sleep propensity Carskadon and De-

.ment, 1979 . From this it is clear that defining andunderstanding sleepiness is possible only thanksto the development of valid and adequate mea-suring instruments based on different operationaldefinitions of the phenomenon.

0167-8760�01�$ - see front matter � 2001 Elsevier Science B.V. All rights reserved.Ž .PII: S 0 1 6 7 - 8 7 6 0 0 1 0 0 1 3 8 - 6

( )G. Curcio et al. � International Journal of Psychophysiology 41 2001 251�263252

Historically, the first distinction was betweenŽsubjective sleepiness as a self-perception of a

. Žhypo-activated state and objective sleepiness asan evaluation of a state of decreased vigilance, onthe basis of specific behavioral or physiological

.parameters . Subsequently, three evaluation lev-els of sleepiness were proposed: the subjective;behavioral-performance; and physiological level.Obviously, different levels of evaluation refer tospecific methods for evaluating and quantifyingsleepiness, but these different methods, as we willsee, very often do not correlate between eachother.

The subjective level has used scales for self-evaluating physical and cognitive symptoms, ask-ing subjects to indicate which definition fromthose proposed matches their own state at the

˚Žmoment of compilation Akerstedt and Gillberg,.1990 .

The behavioral level of sleepiness evaluation,instead, aims to highlight those performance de-

Žcrements for tasks requiring attention Hockey,. Ž .1970 , memory Baddeley, 1968 or general cogni-

Ž .tive effort Wilkinson, 1970 , or coordination ofŽ .motor activity Mullaney et al., 1985 .

The third level, finally, is based on the record-ing of spontaneous variations of several physio-

Žlogical parameters breathing, cardio-vascular,oculomotor, skin conductance and temperature

.parameters at the time of the highest sleeppropensity. Moreover, the same parameter ofsleep propensity has been considered as a perma-nent and lasting tendency to fall asleep: at thispoint they are led to the Multiple Sleep Latency

Ž .Test MSLT; Carskadon and Dement, 1979 . Morerecently, there has been a flourishing of moreelaborate techniques that, for example, consider

˚Žspectral power of EEG Torsvall and Akerstedt,.1987; Stampi et al., 1993 , evoked potentials

Ž .Broughton, 1982 or topographical brain map-ping at the wake�sleep transition and during sleepŽ .Tanaka et al., 2000; Buchsbaum et al., 1982 .

Therefore, following Kamiya’s classificationŽ .Kamiya, 1961 , we can identify four kinds ofmeasures in the field of sleepiness research:

Ž .1. subjective measures self-evaluations ;2. performance decrease measures;

3. sleep propensity measures; and4. arousal decrease measures.

2. Subjective measures or self-evaluations

Subjective measures can be divided into twoŽ .categories: those more numerous that consider

sleepiness as a state characteristic, i.e. as a conse-quence of daily vigilance fluctuations or caused by

Ž .atypical situations i.e. sleep deprivation andthose that, instead, identify daytime sleepinesswith a trait component, i.e. as a steady and con-stant aspect of each subject.

2.1. Sleepiness as a state-related condition

ŽThe Stanford Sleepiness Scale SSS; Hoddes et.al., 1971 belongs to this category. Devised with

the main aim of measuring sleepiness levels inpatients with sleep complaints, the SSS has turnedout to be a useful experimental instrument evenwith normal subjects, and to become the mostlargely used subjective instrument for sleepinessevaluation. The SSS is a Lickert-type scale, withseven progressive levels of vigilance: subjects areasked to indicate which level better describes

Žtheir present state. Many studies Hoddes et al.,.1973; Herscovitch and Broughton, 1981 have

shown that its sensitivity to reveal sleepiness in-creases consequently to periods of sustained sleepdeprivation, while it seems to be not as sensitivewith patients suffering from apnea, who tend to

Ž .deny their trouble e.g. Dement et al., 1978 , orfor distinguishing between apneics and normal

Ž .sleepers Roth et al., 1980 . However, it is con-firmed to be an optimal subjective scale in nar-

Žcolepsy�cataplexy e.g. Aguirre and Broughton,.1987; Broughton and Aguirre, 1987 .

Another subjective scale is the Karolinska˚Ž .Sleepiness Scale KSS created by Akerstedt and

Ž .Gillberg 1990 . This is a nine-step scale: here,too, subjects have to indicate the description-stepthat better reflects the psycho-physical state expe-rienced in the last 10 min. The KSS seems sensi-tive to sleepiness fluctuations: several studies haveshown that KSS scores rise with increased periods

( )G. Curcio et al. � International Journal of Psychophysiology 41 2001 251�263 253

˚Žof sustained wakefulness Akerstedt and Gillberg,.1990 and strongly correlate with time-of-day

˚Ž .Kecklund and Akerstedt, 1993 .˚A second instrument proposed by Akerstedt

and colleagues is the Accumulated Time withŽ .Sleepiness Scale ATS; Gillberg et al., 1994 : this

scale presents eight items and subjects are askedto evaluate the frequency of incidence for severalsymptoms correlated to sleepiness. The ATS hasshown to be sensitive to sleepiness variations in astudy in which it was administered at differenttimes in the night and it seems sensitive also inrevealing different levels of sleepiness evoked byvigilance tasks with varying times on task, dull-

Ž .ness and attentional load Gillberg et al., 1994 .At the same time, there has been increasing

Ž .development of Visual Analogue Scales VASsŽ .for a review see: Monk, 1987 . A VAS is a100-mm long line: subjects have to evaluate theirown state by making a stroke with a pen along theline; the stroke is a measure of sensation inten-sity. Two different kinds of VASs are used in theliterature: unipolar and bipolar. The first typeasks for evaluation of a sensation between the

Žextremes ‘not at all’ and ‘very much’ Monk et al.,.1985; Volk et al., 1994 ; the second, instead, puts

questions like ‘How do you feel now?’ with re-spect to dimensions such as alert�sleepy or ener-

Žgetic�tired Folstein and Luria, 1973; Bond and.Lader, 1974 . VASs are very sensitive to partial or

Ž .total sleep deprivation Casagrande et al., 1997b ,˚Ž .to time of day Akerstedt and Gillberg, 1990 and

Žto the subject’s psycho-physical state Monk,.1987 .

2.2. Sleepiness as a permanent trait component

This category of subjective measures evaluatessleepiness as a trait component, that is, steadyand constant within each subject and uses ques-tionnaires rather than scales. The Epworth

Ž .Sleepiness Scale ESS; Johns, 1991 hypothesizesthat all individuals are characterized by a con-stant level of sleepiness independently of circa-dian or ultradian rhythms: this constant level isclearly evident in one’s likelihood of falling asleepunder low stimulating situations. The ESS askssubjects to indicate their probability of falling

asleep under eight different, typical situations ofdaily life. The ESS seems to be a useful instru-ment in the clinical field for revealing daytimesleepiness: it discriminates between normal andpathological subjects and is sensitive to changesinduced by specific treatments for sleep apneasŽ .Johns, 1991, 1992, 1994 .

Another questionnaire introduced for evaluat-ing sleepiness is the Rotterdam Daytime Sleepi-

Ž .ness Scale RDSS; van Knippenberg et al., 1995 ,created to avoid typical limits of sleep apneadiagnosis. Results show that the RDSS offers a

Žgood convergent and discriminative validity van.Knippenberg et al., 1995 , but it was only used in

the above cited validation study. Therefore, itseems rather excessive to draw conclusions on itsreliability and validity.

Finally, another widely used instrument is theŽSleep�Wake Activity Inventory SWAI; Rosen-

.thal et al., 1993a . The SWAI is a multi-dimen-sional questionnaire composed of 59 items; it isadopted particularly in clinical investigations forevaluating dimensions of physical fatigue, socialdesirability and daytime activity level. Each itemis presented on a nine-step scale and it askssubjects to evaluate the intensity of a given symp-tom. The SWAI seems to be a valid instrumentfor distinguishing between normal subjects andpatients affected by excessive daytime sleepinessŽ .Johnson et al., 1999 .

The SSS was the first self-evaluation instru-ment characterized by easy administration andimmediacy. This characteristic is typical of allsubjective measures, both scales and question-naires-alike. VASs present the further advantageof being easily used for their clear perceptualcomponent. From a psychological and statisticalpoint of view, moreover, visuo-analogic scales arestrongly reliable since they offer answers along acontinuum, without numerical values or ‘locked’

Ž .categories few, very much, enough , that oftenseem forced and artificial.

The main problem for subjective measures re-mains their extreme vulnerability to environmen-tal and motivational variables: very often, individ-uals tend to overestimate their own disorder andso self-evaluations risk giving an exaggerated im-pression of the subject’s psycho-physical state.

( )G. Curcio et al. � International Journal of Psychophysiology 41 2001 251�263254

3. Performance decrease measures

It is well-known that sleepy subjects are oftenimpaired in their perceptual skills, reasoning abil-ities, judgment and decision-making capabilitiesŽ .Dinges and Kribbs, 1991; Bonnet, 1994 . Theeffects of this distortion are clear in an increaseof response omissions, general cognitive slowing,

Ž .memory deficits, increase of reaction times RTs ,etc. In short, the more the brain becomes ‘sleepy’as a result of endogenous factors, the more asubject’s answer will be influenced by context

Ž .factors Dinges, 1987 .Therefore, performance tasks must be divided

into at least two main categories: psychomotorŽ .and cognitive tests see Table 1 . Psychomotor

Žtasks include mainly acoustic and visual RTs sim-.ple or with multiple choice , tracking and tapping

tasks. Cognitive ones, instead, can be divided intoattentional, memory and logical reasoning tasks.

3.1. Psychomotor tasks

The Wilkinson Auditory Vigilance Task iswidely used in the field of vigilance fluctuationsŽ .WAVT; Wilkinson, 1969 . The subject is asked to

Ž .distinguish between ‘brief’ sounds target andŽ .longer ones standard . Administration is dichotic

and to answer one has to push on a button in caseof target tone detection. The length of the WAVTcan vary between 30 and 60 min. Performance is

Table 1Categories of performance measures for sleepiness�vigilanceevaluation

Performance tasks

Psychomotor tests Cognitive tests

Reaction times Attentional tasksAcoustics SubstitutionVisual CancellationSimpleMultiple choice

Tracking tasks Memory tasks

Tapping tasks Logical reasoning tasksArithmetic tasksGrammatical tasks

Ž .evaluated on the basis of hits Hi , omissionsŽ . Ž .OM , false positives FP , RTs to target tones.The WAVT has shown to be sensitive to vigilancefluctuations as a consequence of sleep depriva-tion, but it is also long and boring, and it needscumbersome experimental apparatus. To over-come these limits, Wilkinson and colleagues cre-ated two briefer and easier to administer vigi-lance tasks.

ŽThe Simple Reaction Time Test Glenville et.al., 1978 is a simple RT test of acoustic stimula-

tion and it only lasts 10 min. Subjects have toŽanswer all the presented stimuli presented at

.varying inter-stimuli intervals as soon as possible.However, the Four-Choice Reaction Test

Ž .Wilkinson and Houghton, 1975 is, instead, acomplex RT task. A panel contains four lightsand four buttons, in the same topographical dis-tribution: lights turn on randomly one at a timeand subjects have to push the button correspond-ing to the light that lit up. Data analysis considersRT to hits, errors and omissions.

Other tests have been proposed to assess vigi-lance variations. The Multiple Vigilance TestŽ .MVT; Hishkowitz and Herman, 1989 asks sub-

Žjects to discriminate between visual targets usu-. Žally letters and standard stimuli the same letter

.rotated by 90� . It lasts 30 min and subjects haveto answer as soon as possible. Answer evaluationconsiders hits, omissions, false positives and RTsmean.

ŽThe Simulated Assembly Line Task SALT;.Walsh et al., 1992 is a sustained vigilance task

with a maximum time-on-task of 60 min. Subjectsare asked to sit in front of a computer displayand, by using the mouse, ‘repair’ damages occur-ring to an electronic circuit simulated on themonitor. Dependent variables are even hits, omis-sions and RT.

The tracking task was devised in the neuropsy-chology field in order to evaluate apraxia andneglect syndromes. It consists of using a mouse tofollow a spot of light that moves randomly on acomputer screen. Several studies report that, athigher levels of sleepiness, performance signifi-

Žcantly worsens Webb and Levy, 1982; Mullaney.et al., 1985 .

A peculiar performance decrease instrument is

( )G. Curcio et al. � International Journal of Psychophysiology 41 2001 251�263 255

Žthe Finger-Tapping task Casagrande et al.,.1997a ; based on self-generated motor responses,

it seems sensitive to the sleepiness state and tothe wake-sleep transition. The subject is equippedwith one or two paddles with a button to pushrepetitively and continuously with the forefinger.A dependent variable considered is inter-tapping

Ž .intervals ITI .

3.2. Cogniti�e tasks

Among the widely used cognitive tasks, we canŽconsider the Wilkinson Addition Test Wilkinson,

.1970 . Subjects are asked to add the highest num-ber of groups of five two-figure numbers. Scoresare based on calculation speed and the number oftotal additions; accuracy takes into account thepercentage of correct additions completed in amaximum time-on-task of 30 min. Similararithmetic tasks of forward or backward counting,

Ž .of subtractions Ferrara et al., 2000 or linguisticŽ .ones text understanding, naming are as

Žwidespread in sleepiness research Dinges and.Kribbs, 1991 .

A typical logical reasoning task was proposedŽ .by Baddeley 1968 . The subject is given a sen-

tence like ‘A follows B’ and then, alternatively,patterns like ‘AB’ or ‘BA’. Subjects have to, asquickly as possible, indicate if the pattern pre-sented matches the previously read sentence.

ŽThe Williams Word Memory Test Williams et.al., 1966 is usually composed of a sequence of 25

four-letters words. Words are read at a rate ofone every 10 s; subsequently, subjects are askedto repeat them in any order. Performance is eval-uated on the basis of the number of correctrecalls.

There is also the broad category of attentionaltasks: this kind of test is based on the hypothesisthat, in case of excessive sleepiness, the attentio-nal level drops, thus strongly worsening perfor-mance to several tasks, such as those of visualsearch.

The pioneer of these studies was MackworthŽ .1950 . These tasks were later adopted by cogni-

Ž .tive psychologists, and especially by Neisser 1957to study human information processing.

Another well-defined field of research is that of

substitution tasks. The most used is the Digit-Ž .Symbol Substitution Test DSST , a sub-scale of

the Wechsler�Bellevue Adult Intelligence ScaleŽ .Wechsler, 1958 . It consists of a matrix carrying,on the upper border, nine symbols associated withnine numbers: subjects are asked to match eachsymbol in the matrix with the corresponding num-ber. Task duration is usually of 2 min and it isself-paced. Results take into account total num-ber of attempts and hit percentage.

Another research line looked at detection andcancellation tests, very similar to the one pro-posed in the 1950s by Neisser. Targets used have

Žbeen very different: symbols Gundel and Weg-. Ž .mann, 1989 , numbers Johnson and Diller, 1983 ,

Ž . Žletters Casagrande et al., 1997b , words Webb.and Levy, 1982 . In these tasks, dependent vari-

ables are hits, omissions, false positives, comple-tion time, d’ and ß: the last two parameters arecalculated on the basis of the Signal Detection

Ž .Theory SDT; Tanner and Swets, 1954In conclusion, the specialized literature has

proposed several behavioral methods for survey-ing and evaluating sleepiness both physiologicaland consequent to sleep deprivation. These in-struments present different levels of sensitiveness

Žin evaluating vigilance fluctuations for a review:.Casagrande et al., 1997b , highlighting severalŽ .methodological problems Webb, 1992 : from the

Žchoice of suitable variables Dinges and Kribbs,. Ž1991 , to eminently statistical problems Dinges,. Ž1992 , motivational issues Horne et al., 1983;

.Naitoh, 1983 , and to learning component prob-Ž .lems a review in Millar, 1992 . We may therefore

conclude that behavioral tasks, like subjectiveones, also feel the effects of motivational, envi-ronmental, methodological and statistical prob-lems. However, they present an undoubted advan-tage: they are often the only usable methods infield investigations, and at times are extremelyuseful in experimental studies, too, because theyallow a considerable reduction in the intrusivenature of the measure.

4. Sleep propensity measures

The most famous and widespread method for

( )G. Curcio et al. � International Journal of Psychophysiology 41 2001 251�263256

an objective evaluation of sleepiness is the Multi-Ž .ple Sleep Latency Test MSLT , devised by

Ž .Carskadon and Dement 1979 . It is based on theidea that sleepiness is a state in which propensityto falling asleep is particularly urgent. It is possi-ble, therefore, to quantify the subject’s sleepinesslevel by polygraphically measuring the speed offalling asleep or sleep latency.

Sleep latency is calculated starting from lightsoff to the first of three consecutive sleep epochsŽstate 1, 2, 3, 4, REM, following Rechtschaffen

.and Kales rules; Rechtschaffen and Kales, 1968 .If subjects do not fall asleep the maximum dura-tion of the trial is 20 min: after this period, theycan get up and continue their activities. Otherpolygraphic parameters have been proposed forevaluating sleepiness. They are not limited toevaluating falling asleep propensity but also con-sider the architecture of the following sleep stages,to provide a more complete description of diurnalsleep propensity. Thus the Polygraphic Index of

Ž .Sleepiness PIS and Polygraphic Score of Sleepi-Ž . Ž .ness PSS Roth et al., 1984, 1986 have been

developed. With respect to standard MSLT thefalling asleep trial is unique and usually placedbetween 14.00 and 16.00 h: scores are based on apondered measure derived from sleep latency,sleep stage latencies and duration of the fol-lowing sleep stages. In this way, the authors reli-ably discriminated between patients affected bysleep disorders and control subjects.

Several studies have indicated the MSLT as themost sensitive to sleepiness variations during 24 hŽ .Richardson et al., 1982; Levine et al., 1988 , total

Žsleep deprivation Carskadon and Dement, 1979;.Carskadon et al., 1981 , partial sleep deprivation

Ž .Rosenthal et al., 1993b and to sleep extensionŽ .Carskadon and Dement, 1987 . Moreover, the

ŽMSLT is also widely used in clinics Report from.the American Sleep Disorders Associations, 1992 ,

where, nevertheless, it has sometimes shown re-duced sensitivity: while it seems a good method indiagnoses of narcolepsy and sleep apneas, it alsoprovides contrasting data on insomnias anddaytime sleepiness evaluations consequent onpharmacological treatments.

To avoid a ‘floor effect’ typical of patientsŽ .suffering from excessive daytime sleepiness EDS ,

showing very low latencies independently of thereal level of sleepiness, other sleep propensitytests have been proposed. Since EDS patientssubmitted to pharmacological treatments im-proved only in their ability to remain awake, butdid not show a decrease of sleep propensity, alter-native tests to the MSLT, assessing the ability ofthe subject to resist sleep, have been developed.The Repeated Test of Sustained WakefulnessŽ .RTSW; Hartse et al., 1982 , was thus devised,which is similar to the MSLT with the exceptionof requests: here subjects have to stay awakealthough in bed and with lights off. It has beenobserved that instruction is important, becauseMSLT and RTSW studies have noticed signifi-cantly higher sleep latencies in second testsŽ .Hartse et al., 1982 .

An alternative to the RTSW is the Mainte-Žnance of Wakefulness Test MWT; Mitler et al.,

.1982 . In this case subjects have to avoid fallingasleep while comfortably seated on an armchairand in a non-stimulating situation. Also, this testseems useful in evaluating pharmacological treat-

Ž .ments of narcoleptics Mitler et al., 1982 .Subsequently, a cross-validation study was car-

ried out using the MSLT, RTSW and MWTŽ .Sugerman and Walsh, 1989 . Results againshowed that the MSLT is the most sensitivemethod for normal subjects, while the RTSW andMWT are better when one wants to evaluatesubjects characterized by extreme levels of sleeppropensity and when one needs differential diag-noses of EDS problems. However, recently, JohnsŽ .2000 , raised some doubts about the real sensi-tiveness of the MSLT, showing that it fails to bethe gold standard in daytime sleepiness evalua-tion, probably because, firstly, it is a measure onlyof a situational sleep propensity and, secondly,because it does not clearly distinguish betweennormal and abnormal daytime sleepiness levels,adopting rules derived from subjective ‘impres-

Žsions’ and not from clinical evidence Richardson.et al., 1978 .

In conclusion, we can say that measures ofsleep propensity are clearly optimal tests forsleepiness evaluation, both in research and verygood in clinics. It is evident that the variousillustrated tests are very good clinical tests useful

( )G. Curcio et al. � International Journal of Psychophysiology 41 2001 251�263 257

in diagnosing several sleep pathologies, from EDSdisorders to narcolepsy and insomnia. Limits arelinked to invasivity of techniques and to instru-ments necessary for recording; such limits areeasily accepted in clinics but are clearly obstaclesin research studies. Moreover, if we accept Johns’observation, we have to conclude that these kindsof measures are acceptable only in clinical inves-tigation, since, as seen previously, the MSLT orits alternatives are only measures of one, specific

Ž .situational sleep propensity Johns, 2000 .

5. Arousal decrease measures

Sleep latency seems to be an optimal parame-ter of sleepiness level, but characteristics of the

ŽMSLT multiple laboratory recordings at several.times during the day are a critical limit to using

this test in different situations, particularly inapplied and field studies. From this and from theattempt to find more accurate measures of sleepi-ness, there arose the interest of some researchersfor the identification of physiological parameters,with the aim of indicating a sleepiness state inawake and actively operating subjects.

The parameters proposed by researchers canbe divided into three main categories: electroen-

Ž .cephalographic EEG parameters, electroculo-Ž .graphic EOG parameters and autonomic indica-

tors.

5.1. EEG parameters

Usually, EEG parameters can be summarizedin EEG patterns, spectral characteristics of the

Ž .EEG and evoked potentials EP .Some authors observed that the presence of

Ž . Ž .alpha 8�12 Hz and theta 4�8 Hz rhythms inthe EEG of awake and active subjects can pro-vide information on the psychophysiological state

Žof sleepiness or lowered vigilance O’Hanlon and.Beatty, 1977; Fruhsthorfer et al., 1977 . Taking

these observations into account, several sleepi-ness-sensitive instruments have been developed,based on the detection of alpha and theta band

˚Žpower increases Torsvall and Akerstedt, 1987,.1988 . One of them is the Alpha Attenuation Test

Ž .AAT; Stampi et al., 1993 , that starts from theexperimentally demonstrated hypothesis thatwhenever an awake individual falls asleep, thespectral power of alpha decreases with eyes closedand increases with eyes open. Therefore, the AATprocedure asks subjects to open and close eyes

Ževery 2 min for every trial total duration: 12.min . Data collected allows working out a vigi-

lance index defined as an ‘alpha attenuation in-Ž .dex’ AAI , given by the ratio between spectral

power with eyes closed and spectral power witheyes open: the lower the AAI, the higher thesleepiness level and vice versa. The AAT hasshown its sensitivity to vigilance decreases fol-

Žlowing one sleep deprivation night Stampi et al.,.1995a and seems to correlate significantly with

Žsleep latencies measured by the MSLT Stampi et.al., 1993 . Moreover, the AAT is a brief practical

diagnostic instrument for evaluating EDS associ-Ž .ated with narcolepsy Alloway et al., 1997 .

The analysis of alpha and theta band spectralpower has also been used in some field studies.Alpha power density, intended as a sleepinessindex, seems to be considerably higher in a nightworkers’ group with respect to an afternoon

˚Ž .workers’ group Akerstedt et al., 1991 , and indiurnal train drivers with respect to night shift

˚Ž .ones Torsvall and Akerstedt, 1987 . A follow-up˚ Ž .study by Torsvall and Akerstedt 1988 indicated

that spectral power of alpha and theta bandsincreases significantly approximately 20 s beforean episode of head falling back, that does nothappen before an answer to a behavioral test.Spectral power of alpha and theta rhythms, more-over, has shown its sensitivity to sleepiness in-crease, evaluated during a sleepless night, with areduced vigilance peak between 05.00 and 09.00 h˚Ž .Akerstedt and Gillberg, 1990 .

Brain mapping techniques have greatly beenŽincreasing in the last few decades e.g. Buchs-

baum et al., 1982; Cantero et al., 1999; Tanaka et.al., 2000 . Tanaka and co-workers, for example,

considered EEG data and coherence of delta andtheta bands, and found that the dominant syn-chronous component of slow wave activity in-creases as a function of EEG stage inanterior�central areas, and only after the appear-ance of vertex sharp waves. These data are not in

( )G. Curcio et al. � International Journal of Psychophysiology 41 2001 251�263258

contrast with the ones proposed previously byŽ .Buchsbaum and colleagues 1982 , and show how

the power density of EEG bands follows a specifictopographical and temporal sequence: this, ofcourse, opens up interesting developments in both

Žresearch and clinical applications Cantero et al.,.1999 .

Another measure based on EEG characteris-tics and used for evaluating vigilance variations

Ž .are evoked potentials EPs . If sleepiness is de-fined as a state that inhibits or at least reducessubject ability to respond to environmental sti-muli, EP latencies will increase and amplitudeswill decrease with respect to a baseline condition.Several studies with normal subjects provided dataconfirming the hypothesis that vigilance decreases

Žonly affect long latency components of EPs N1,.N2 or P2 as well as P300 , inducing a significant

reduction in amplitude and a lengthening ofŽlatency Weitzman and Kremer, 1965; Fruh-

sthofer and Bergstrom, 1969; Wesensten and Ba-.dia, 1988 . In clinics, instead, evoked potentials

are usually used as a diagnostic instrument fornarcolepsy, considering CNV and P300 compo-

Žnents as dependent variables Aguirre and.Broughton, 1987; Broughton and Aguirre, 1987

with a clear sensitivity only for the latter as ameasure of sleepiness.

5.2. EOG parameters

A large number of studies have directed atten-tion to the possibility of considering oculomotoractivity parameters as a measure of sleepiness.This condition of hypovigilance, in fact, is charac-terized by some oculomotor activity: disappear-

Ž .ance of rapid eye movements saccades , reduc-tion of spontaneous blinks, appearance of slow

Ž .eye movements SEMs that increase during thesleep onset period and diminish drastically during

Žsleep Santamaria and Chiappa, 1987; Morris and.Mitler, 1996 . More clearly, saccades present

a positive correlation between speed and sleepŽ .latency Stampi et al., 1995b ; speed, moreover,

turns out to be a reliable measure of deactivationŽ .Galley, 1989 . SEMs, instead, have shown a posi-tive relationship with the spectral power of alpha

˚Žand theta bands Torsvall and Akerstedt, 1987,

.1988 , a negative relation with performance to˚Ž .behavioral tasks Torsvall and Akerstedt, 1988 ,

and agree with subjective ratings of sleepiness˚Ž .Akerstedt and Gillberg, 1990 . Finally, with

diminished blink frequency, a worse performanceto acoustic tasks for evaluating nighttimes vigi-

˚Ž .lance is observed Torsvall and Akerstedt, 1988and to tracking tasks performed during a night-

Ž .time prolonged activity Peled and Lavie, 1994 .Frequency of spontaneous blinks also correlatesnegatively with the total amount of hours of sleep

Ž .deprivation Gray et al., 1995 .

5.3. Autonomic indicators

Finally, variations in some autonomic parame-ters have been evaluated, and the most commontechnique in the field of sleepiness is pupillome-try. This technique has been mainly applied toclinics for diagnosis and quantification of vigi-

Ž .lance disorders typical of EDS Schmidt, 1982 .Pupil dilation and contraction are governed by

Ž .the Autonomic Nervous System ANS and varyŽwith fatigue and sleepiness Lowenstein and

.Lowenfeld, 1958; Wilhelm et al., 1998 . Basically,this technique allows an analysis of pupil diame-ter variations in response to both visual andacoustic stimuli. The use of pupillometry is basedon the observation that a broad and steady pupildiameter is typical of a normal level of vigilance,while a contracted and changeable pupil is at-tributable to sleepiness and hypoactivation condi-

Ž .tions Lavie, 1979 . But these results are contro-versial and, in spite of the above data, some otherstudies demonstrate that pupillometry cannot forthe present be considered as a clearly reliable

Žtechnique of sleepiness evaluation Newman and.Broughton, 1991; Ranzjin and Lack, 1997 .

6. Conclusions

Every conclusion on the validity of sleepinessevaluation methods can be discussed in the lightof three main factors: the different sensitivity ofseveral available tests to evaluate sleepiness lev-els; the influence of sleep debt on sleepinessquantification, and the typical characteristic ofsubjective and performance measures to be vul-

( )G. Curcio et al. � International Journal of Psychophysiology 41 2001 251�263 259

Ž .nerable to external environmental and internalŽ .motivational factors.

Different instruments seem to be differentlysensitive to diurnal sleepiness, both physiologicaland pathological.

ŽMSLT and subjective ratings i.e. SSS, KSS,.ATS show a good sensitivity in revealing daytime

sleepiness in normal subjects, while they arepoorly sensitive with pathological populations.SSS, instead, is not advisable with patients suffer-ing from sleep apnea because they tend to under-

Žestimate or even deny their problem Roth et al.,.1986 , while MSLT showed a floor-effect: the

range of variation with patients is very shortened,Žnot permitting a differential diagnosis Hartse et

.al., 1982 . This has led to the development ofalternative methods such as RTSW, MVT orMWT. Several subjective ratings, instead, haveshown their sensitivity from a diagnostic point ofview: SWAI seems to correlate and even reliablypredict sleep latencies measured using MSLT with

Ž .ESD patients Rosenthal et al., 1993b ; ESSŽ .Johns, 1994 has proved to be a valid and reliabletrait measure of daytime sleepiness, not influ-

Ženced by contingent factors; RDSS, finally van.Knippenberg et al., 1995 , has proved to be a

useful diagnostic instrument for EDS syndromes.VASs, instead, because they are characterized byimmediacy and easiness of completion, are allreliable and valid measures of sleepiness

˚ŽCasagrande et al., 1997b; Akerstedt and Gillberg.1990; Monk, 1987 , both in normal and clinical

populations.Performance tasks have shown great reliability:

Žboth psychomotor Wilkinson, 1969; Casagrande. Žet al., 1997b and cognitive Baddeley, 1968; An-.gus et al., 1985 measures are valid instruments in

the field of sleepiness. Several problems exist forthis kind of test, and involve methodologicalŽ . Ž .Webb, 1992 , motivational Williams et al., 1959

Ž .and learning aspects Millar, 1992 , but none ofthese seem to be insurmountable. Moreover, thecontroversy between using long, complex andtask-paced measures and brief, easy and self-

Ž .paced ones Wilkinson, 1965 , seems to be solvedfrom a methodological point of view, on the basis

Žof clear and unambiguous experimental data for.a review see Dinges, 1992 .

With regard to physiological measures, AATand EOG parameters have shown a good sensitiv-ity to sleepiness, both with normal and pathologi-

˚Žcal subjects Akerstedt and Gillberg, 1990; Al-.loway et al., 1997 . Both techniques prove to be

very promising in field investigations, in that theydo not need complex procedures and tools, andseems to be true ‘ecological’ measures. Evoked

Žpotentials particularly P300 and, generally, long-.latency components have been widely used in

Ždiagnostics of sleep medicine Broughton and.Aguirre, 1987 . Pupillometry has yielded contro-

versial data and its reliability, for the moment, isŽstill unclear Ranzjin and Lack, 1997; Wilhelm et

.al., 1998 .Another problem is the different variability of

several measures of daytime sleepiness as a func-tion of the total amount of prior sleep. Subjective,behavioral and physiological measures agree insleepiness level evaluation in cases of extreme

Ž .sleep deprivation total sleep deprivation , i.e.when a very high level of sleepiness is induced.Several differences in evaluating sleepinessemerge in cases of sleep manipulation paradigmsin which sleepiness levels elicited are lighter.Thus, when we are interested in quantifyingdaytime physiological sleepiness or that followingpartial sleep deprivation, we can observe thatsubjective and behavioral measures are in greateragreement with respect to both behavioral andphysiological measures, on the one hand, and

Žphysiological and subjective measures Dinges and.Kribbs, 1991 , on the other.

ŽUnderstanding this phenomenon or one of the.possible interpretations brings us to the last issue

in sleepiness measuring. Carskadon and DementŽ .1979, 1987 developed the hypothesis that subjec-tive measures correlate significantly with only thebehavioral ones, because both are characterizedby the fact that they are easily influenced bymotivational confounding variables: on the otherhand, measures such as sleep latencies on MSLT,spectral power of EEG bands, evoked potentialsor EOG characteristics are not influenced bysuch biases. Besides, it is well-known that both

Žeffects of task influence final performance John-.son, 1982 , and that objective indices do not cor-

relate with performance or with mood levels, con-

( )G. Curcio et al. � International Journal of Psychophysiology 41 2001 251�263260

versely to what happens with subjective ratingsŽ .Johnson et al., 1991 . These results would con-firm the hypothesis of a greater vulnerability toexogenous and motivational factors of subjectiveand behavioral measures. This hypothesis hasbeen repeatedly proposed by several researchersin different periods offering considerably differ-ent tasks as examples: such as the lapse hypothe-

Ž .sis of Williams and colleagues 1959 or the con-textual dependence hypothesis of Dinges and

Ž .Powell 1989 .

References

Aguirre, M., Broughton, R.J., 1987. Complex event relatedŽ .potentials P300 and CNV and MSLT in the assessment of

excessive daytime sleepiness in narcolepsy�cataplexy. Elec-troencephogr. Clin. Neurophysiol. 67, 298�318.

Akerstedt, T., 1995. Work hours, sleepiness and underlyingŽ .mechanisms. J. Sleep Res. 4 Suppl. 2 , 15�22.

Akerstedt, T., Gillberg, M., 1990. Subjective and objectivesleepiness in the active individual. Int. J. Neurosci. 52,29�37.

Akerstedt, T., Kecklund, G., Knuttson, A., 1991. Manifestsleepiness and the spectral content of the EEG during shiftwork. Sleep 14, 221�225.

Aldrich, M.S., 1994. Parkinsonism. In: Kryger, M.H., Roth, T.,Ž .Dement, W.C. Eds. , Principles and Practice of Sleep

Medicine, 2nd edn Saunders, Philadelphia.Alloway, C.E.D., Ogilvie, R.D., Shapiro, C.M., 1997. The alpha

attenuation test: assessing excessive daytime sleepiness innarcolepsy�cataplexy. Sleep 20, 258�266.

Angus, R.G., Heslegrave, R.J., Myles, W.S., 1985. Effects ofprolonged sleep deprivation, with and without chronic phys-ical exercise, on mood and performance. Psychophysiology22, 276�282.

Baddeley, A., 1968. A 3 minute reasoning task based ongrammatical transformation. Psychon. Sci. 10, 341�342.

Bond, A., Lader, M., 1974. The use of analogue scales inrating subjective feelings. Br. J. Med. Psychol. 47, 211�218.

Bonnet, M.H., 1994. Sleep deprivation. In: Kryger, M.H.,Ž .Roth, T., Dement, W.C. Eds. , Principles and Practice of

Sleep Medicine, 2nd edn Saunders, Philadelphia.Broughton, R., 1982. Performance and evoked potential mea-

sures of various states of daytime sleepiness. Sleep 5,S135�S146.

Broughton, R.J., 1989. Sleep attacks, naps, and sleepiness inmedical sleep disorders. In: Dinges, D.F., Broughton, R.J.Ž .Eds. , Sleep and Alertness, Raven Press, New York.

Broughton, R.J., Aguirre, M., 1987. Differences between REMand NREM sleepiness measured by event-related poten-

Ž .tials P300 and CNV , MSLT and subjective estimate innarcolepsy�cataplexy. Electroencephogr. Clin. Neurophys-iol. 67, 317�326.

Buchsbaum, M.S., Mendelsohn, W.B., Duncan, W.C., Cop-pola, R., Kelsoe, J., Gillin, J.C., 1982. Topographical corti-cal mapping of EEG sleep stages during afternoon naps innormal subjects. Sleep 5, 248�255.

Cantero, J.L., Atienza, M., Salas, R.M., Gomez, C., 1999.Cerebral mapping during sleep: a critical review of litera-ture. Rev. Neurol. 28, 417�423.

Carskadon, M.A., Dement, W.C., 1979. Effects of total sleeploss on sleep tendency. Percept. Mot. Skills 48, 495�506.

Carskadon, M.A., Dement, W.C., 1987. Daytime sleepiness:quantification of a behavioral state. Neurosci. Biobehav.Rev. 11, 307�317.

Carskadon, M.A., Harvey, K., Dement, W.C., 1981. Sleep lossin young adolescents. Sleep 4, 299�312.

Casagrande, M., De Gennaro, L., Violani, C., Braibanti, P.,Bertini, M., 1997a. A finger-tapping task and a reactiontime task as behavioral measures of the transition fromwakefulness to sleep: which task interferes less with thesleep onset process? Sleep 20, 301�312.

Casagrande, M., Violani, C., Curcio, G., Bertini, M., 1997b.Assessing vigilance through a brief pencil and paper letter

Ž .cancellation task LCT : effects of one night of sleep depri-vation and of the time of day. Ergonomics 40, 613�630.

Dement, W.C., Carskadon, M.A., Richardson, G., 1978. Exces-sive daytime sleepiness in the sleep apnea syndrome. In:

Ž .Guilleminault, C., Dement, W.C. Eds. , Sleep Apnea Syn-dromes, Alan R. Liss Inc, New York.

Dement, W.C., Gelb, M., 1993. Somnolence: its importance insociety. Neurophysiol. Clin. 23, 5�14.

Dinges, D.F., 1987. The nature of sleepiness: causes, contextŽ .and consequences. In: Strunkard, A., Baum, A. Eds. ,

Eating, Sleeping and Sex, Academic Press, New York.Dinges, D.F., 1992. Probing the limits of functional capacity:

the effects of sleep loss on short duration task. In: Ogilvie,Ž .R., Broughton, R. Eds. , Sleep, Arousal and Performance,

Birkhauser, Boston.Dinges, D.F., 1995. An overview of sleepiness and accidents. J.

Ž .Sleep Res. 4 Suppl. 2 , 4�14.Dinges, D.F., Kribbs, N.B., 1991. Performing while sleepy:

effects of experimentally-induced sleepiness. In: Monk, T.H.Ž .Ed. , Sleep, Sleepiness and Performance, John Wiley andSons, Chichester.

Dinges, D.F., Powell, J., 1989. Sleepiness impairs optimumresponse capability � it’s time to move beyond the LapseHypothesis. Sleep Res. 18, 366.

Ferrara, M., De Gennaro, L., Casagrande, M., Bertini, M.,2000. Selective slow-wave sleep deprivation and time-of-night effects on cognitive performance upon awakening.Psychophysiology 37, 440�446.

Folstein, M.F., Luria, R., 1973. Reliability, validity and clinicalapplication of the visual analogue mood scale. Psychol.Med. 3, 479�486.

Fruhsthorfer, H., Bergstrom, R., 1969. Human vigilance and

( )G. Curcio et al. � International Journal of Psychophysiology 41 2001 251�263 261

auditory evoked responses. Electroencephogr. Clin. Neuro-physiol. 27, 346�355.

Fruhsthorfer, H., Langanke, P., Meinzer, K., Peter, J.H., Pfaff,U., 1977. Neurophysiological vigilance indicators and oper-ational analysis of a train vigilance monitoring device: a

Ž .laboratory and field study. In: Mackie, R.R. Ed. , Vigi-lance, Plenum Press, New York.

Galley, N., 1989. Saccadic eye movement velocity as an indica-Ž .tor of de -activation. A review and some speculations. J.

Psychophysiol. 3, 229�244.˚Gillberg, M., Kecklund, G., Akerstedt, T., 1994. Relations

between performance and subjective ratings of sleepinessduring a night awake. Sleep 17, 236�241.

Glenville, M., Broughton, R., Wing, A.M., Wilkinson, R.T.,1978. Effects of sleep deprivation on short duration perfor-mance measures compared to the Wilkinson auditory vigi-lance task. Sleep 1, 169�176.

Gray, P.W., Allen, R.P., Thomas, M.L., Thorne, D.R., Sing,H.C., Quigg Davis, H., 1995. Eye blink rates during asimulated driving task before and after sleep deprivation.Sleep Res. 24, 445.

Gundel, A., Wegmann, H.M., 1989. Transition between ad-vance and delay responses to eastbound transmeridianflight. Chron. Int. 6, 147�156.

Hartse, K.M., Roth, T., Zorick, F.J., 1982. Daytime sleepinessand daytime wakefulness: the effect of instruction. Sleep 5,S107�S118.

Herscovitch, J., Broughton, R., 1981. Performance deficitsfollowing short-term partial sleep deprivation and subse-quent recovery oversleeping. Can. J. Psychol. 35, 309�322.

Hishkowitz, M., Herman, J., 1989. Multiple vigilance testŽ .MVT . Sleep Res. 18, 393.

Hockey, G.R.J., 1970. Changes in attention allocation in amulti-component task under loss of sleep. Br. J. Psychol.61, 473�480.

Hoddes, E., Dement, W.C., Zarcone, V., 1971. The history anduse of the Stanford Sleepiness Scale. Psychophysiology 9,150.

Hoddes, E., Zarcone, V., Smythe, H., Phillips, R., Dement,W.C., 1973. Quantification of sleepiness: a new approach.Psychophysiology 10, 431�436.

Horne, J.A., Anderson, N.R., Wilkinson, R.T., 1983. Effects ofsleep deprivation on signal detection measures of vigilance:implications for sleep function. Sleep 6, 347�358.

Horne, J.A., Reyner, L.A., 1995. Driver sleepiness. J. SleepŽ .Res. 4 Suppl. 2 , 23�29.

Johns, M.W., 1991. A new method for measuring daytimesleepiness: the Epworth Sleepiness Scale. Sleep 14, 540�545.

Johns, M.W., 1992. Reliability and factor analysis of theEpworth Sleepiness Scale. Sleep 15, 376�381.

Johns, M.W., 1994. Sleepiness in different situations measuredby the Epworth Sleepiness Scale. Sleep 17, 703�710.

Johns, M.W., 2000. Sensitivity and specivity of the multipleŽ .sleep latency test MSLT , the maintenance of wakefulness

test and the Epworth sleepiness scale: failure of the MSLTas a gold standard. J. Sleep Res. 9, 5�11.

Johnson, E.O., Breslau, N., Roth, T., Roehrs, T., Rosenthal,L., 1999. Psychometric evaluation of daytime sleepiness andnocturnal sleep onset scales in a representative communityscales. Biol. Psychol. 45, 764�770.

Johnson, L.C., 1982. Sleep deprivation and performance. In:Ž .Webb, W.B. Ed. , Biological Rhythms, Sleep and Perfor-

mance, Wiley, Chichester.Johnson, L.C., Freeman, C.R., Spinweber, C.L., Gomez, S.A.,

1991. Subjective and objective measures of sleepiness: ef-fect of benzodiazepine and caffeine on their relationship.Psychophysiology 28, 65�71.

Johnson, C.W., Diller, L., 1983. Error evaluation ability ofright-hemisphere brain-lesioned patients who have hadperceptual�cognitive retraining. J. Clin. Neuropsychol. 5,401�402.

Kamiya, J., 1961. Behavioral, subjective and physiological as-pects of drowsiness and sleep. In: Fiske, D.W., Maddi, S.R.Ž .Eds. , Functions of Varied Experience, Dorsey Press,Homewood.

˚Kecklund, G., Akerstedt, T., 1993. Sleepiness in long distancetruck driving: an ambulatory EEG study of night driving.Ergonomics 36, 1007�1117.

Kleitman, N., 1963. Sleep and Wakefulness, The University ofChicago Press, Chicago.

Lavie, P., 1979. Ultradian rhythm in alertness. A pupillometricstudy. Biol. Psychol. 9, 49�62.

Levine, B., Rohers, T., Zorick, F., Roth, T., 1988. Daytimesleepiness in young adults. Sleep 11, 39�46.

Lowenstein, O., Lowenfeld, I., 1958. Electronic pupullography� a new instrument and some clinical application. Arch.Ophtalm. 59, 352�363.

Mackworth, J.F., 1950. Researches on the measurement ofhuman performance, Medical Research Council SpecialReport No. 268. H.M., Stationary Office, London.

Millar, K., 1992. Some chronic problems in the methodologyof performance tasks applied in clinical settings. In: Ogilvie,

Ž .R., Broughton, R. Eds. , Sleep, Arousal and Performance,Birkhauser, Boston.

Mitler, M.M., Gujavarty, K.S., Browman, C.P., 1982. Mainte-nance of wakefulness test: a polysomnographic techniquefor evaluating treatment efficacy in patients with excessivesomnolence. Electroencephogr. Clin. Neurophysiol. 53,658�666.

Monk, T.H., 1987. Subjective ratings of sleepiness � theunderlying circadian mechanisms. Sleep 10, 343�353.

Monk, T.H., Fookson, J.E., Moline, M.L., Pollak, C.P., 1985.Diurnal variation in mood and performance in a time-iso-lated environment. Chronobiol. Int. 2, 185�293.

Morris, T.L., Mitler, J.C., 1996. Electrooculographic and per-formance indices of fatigue during simulated flight. Biol.Psychol. 42, 343�360.

Mullaney, D.J., Fleck, P.A., Okudaira, N., Kripke, D.F., 1985.An automated system for administering continuous work-load and for measuring sustained continuous performance.Behav. Res. Meth. Instr. Comp. 17, 16�18.

( )G. Curcio et al. � International Journal of Psychophysiology 41 2001 251�263262

Naitoh, P., 1983. Signal detection theory as applied to vigi-lance performance of sleep-deprived subjects. Sleep 6,359�361.

Neisser, U., 1957. Cognitive Psychology, Appleton-Century-Crofts, New York.

Newman, J., Broughton, R., 1991. Pupillometric assessment ofexcessive daytime sleepiness in narcolepsy�cataplexy. Sleep14, 121�129.

O’Hanlon, J.F., Beatty, J., 1977. Concurrence of electroen-cephalographic and performance changes during a simu-lated radar watch and some implications for the arousal

Ž .theory of vigilance. In: Mackie, R.R. Ed. , Vigilance,Plenum Press, New York.

Peled, N., Lavie, P., 1994. Monitoring oprator’s performanceŽ .� the role of blinking. J. Sleep Res. 3 Suppl. 1 , 196.

Porcu, S., Casagrande, M., Ferrara, M., Bellatreccia, A., 1998.`Sleep and alertness during alternating monophasic andpolyphasic rest�activity cycles. Int. J. Neurosci. 95, 43�50.

Ranzjin, R., Lack, L., 1997. The pupillary light reflex cannotbe used to measure sleepiness. Psychophysiology 34, 17�22.

Rechtschaffen, A., Kales, A., 1968. A Manual of StandarizedTerminology, Techniques, and Scoring System for SleepStages of Human Subjects, BIS�BRI, UCLA, Los Angeles.

Report from the American Sleep Disorders Associations, 1992.The clinical use of the Multiple Sleep Latency Test. Sleep15, 268�276.

Richardson, G.S., Carskadon, M.A., Flagg, W., van den Hoed,J., Dement, W.C., Mitler, M.M., 1978. Excessive daytimesleepiness in man: multiple sleep latency measurement innarcoleptic and control subjects. Electroencephogr. Clin.Neurophysiol. 45, 621�627.

Richardson, G.S., Carskadon, M.A., Orav, E.J., Dement, W.C.,1982. Circadian variation of sleep tendency in elderly andyoung adult subject. Sleep 5, S82�S92.

Rosenthal, L., Roehrs, T., Roth, T., 1993a. The sleep�wakeactivity inventory: a self-report measure of daytime sleepi-ness. Biol. Psychiat. 34, 810�820.

Rosenthal, L., Roehrs, T.A., Rosen, A., Roth, T., 1993b. Levelof sleepiness and total sleep time following various time inin bed conditions. Sleep 16, 226�232.

Roth, T., Hartze, K.M., Zorick, F., Conway, W., 1980. Multiplenaps and the evaluation of daytime sleepiness in patientswith upper airway sleep apnea. Sleep 3, 425�439.

Roth, T., Nemsivalova, S., Sonka, K., Docekal, P., 1984. Aquantitative poligraphic study of daytime somnolence andsleep in patients with excessive diurnal sleepiness. Arch.Suisse Neurol. Neurochir. Psychiat. 135, 265�272.

Roth, T., Nemsivalova, S., Sonka, K., Docekal, P., 1986. Analternative to the Multiple Sleep Latency Test for de-termining sleepiness in narcolepsy and hypersomnia: apolygraphic score of sleepiness. Sleep 9, 243�245.

Santamaria, J., Chiappa, K.H., 1987. The EEG of drowsinessin normal adults. J. Clin. Neurophysiol. 4, 327�382.

Schmidt, H.S., 1982. Pupillometric assessment of disorders ofarousal. Sleep 5, S157�S164.

Stampi, C., Stone, P., Michimori, A., 1993. The alpha attenua-tion test: a new quantitative method for assessing sleepi-ness and its relationship to the MSLT. Sleep Res. 22, 115.

Stampi, C., Stone, P., Michimori, A., 1995a. A new quantita-tive method for assessing sleepiness: the alpha attenuationtest. Work Stress 9, 368�376.

Stampi, C., Aguirre, A., Macchi, M., Hashimoto, S., 1995b.Relationship between pupillary�oculomotor measures, and

Ž .physiological EEG and behavioral indices of sleepiness.Sleep Res. 24A, 454.

Sugerman, J.L., Walsh, J.K., 1989. Physiological sleep ten-dency and ability to maintain alertness at night. Sleep 12,106�112.

Tanaka, H., Hayashi, M., Hori, T., 2000. Topographical char-acteristics of slow wave activities during the transition fromwakefulness to sleep. Clin. Neurophysiol. 111, 417�427.

Tanner, W.P., Swets, J.A., 1954. A decision-making theory ofvisual detection. Psychol. Rev. 61, 401�409.

˚Torsvall, L., Akerstedt, T., 1987. Sleepiness on the job: contin-uously measured EEG changes in train drivers. Electroen-cephogr. Clin. Neurophysiol. 66, 502�511.

˚Torsvall, L., Akerstedt, T., 1988. Extreme sleepiness: quanti-fication of EOG and spectral EEG parameters. Int. J.Neurosci. 38, 435�441.

van Knippenberg, F.C.E., Passchier, J., Heysteck, D. et al.,1995. The Rotterdam Daytime Sleepiness Scale: a newdaytime sleepiness scale. Psycholog. Rep. 76, 83�87.

Volk, S., Dyroff, J., Georgi, K., Pflug, B., 1994. Subjectivesleepiness and physiological sleep tendency in healthy youngmorning and evening subject. J. Sleep Res. 3, 138�143.

Walsh, J.K., Muelbach, M.J., Schweitzer, P.K., 1992. Simulatedassembly line performance following ingestion of cetirizineor hydroxyzine. Ann. Allerg. 69, 195�200.

Webb, W.B., 1992. Problems in measuring performance: dan-gers in difference scores. In: Ogilvie, R., Broughton, R.Ž .Eds. , Sleep, Arousal and Performance, Birkhauser, Bos-ton.

Webb, W.B., Levy, C.M., 1982. Age, sleep deprivation, andperformance. Psychophysiology 19, 272�276.

Wechsler, D., 1958. The Measurement and Appraisal of AdultIntelligence, 4th edn Williams Wilkins, Baltimora.

Weitzman, E.D., Kremer, H., 1965. Auditory evoked responsesduring different stages of sleep in man. Electroencephogr.Clin. Neurophysiol. 18, 65�70.

Wesensten, N.J., Badia, P., 1988. The P300 component insleep. Physiol. Behav. 44, 215�220.

Wilhelm, B., Whilelm, H., Ludtke, H., Streicher, P., Adler, M.,¨1998. Pupillographic assessment of sleepiness in sleep de-prived healthy subjects. Sleep 21, 258�265.

Wilkinson, R.T., 1965. Sleep deprivation. In: Edholm, O.G.,Ž .Bacharach, A.L. Eds. , The Physiology of Human Survival,

Academic Press, London.Wilkinson, R.T., 1969. Some factors influencing the effect of

environmental stresses upon performance. Psychol. Bull.72, 260�272.

( )G. Curcio et al. � International Journal of Psychophysiology 41 2001 251�263 263

Wilkinson, R.T., 1970. Methods for research on sleep depriva-Ž .tion and sleep function. In: Hartmann, H. Ed. , Sleep and

Dreaming, Little Brown, Boston.Wilkinson, R.T., Houghton, D., 1975. Portable four-choice

reaction time test with magnetic tape memory. Behav. Res.Meth. Instr. Comp. 7, 441�446.

Williams, H., Gieseking, C., Lubin, A., 1966. Some effects ofsleep loss on memory. Percept. Mot. Skills 23, 1287�1293.

Williams, H.L., Lubin, A., Goodnow, J.J., 1959. Impairedperformance with acute sleep loss. Psychol. Monogr. 73,1�26.