How to measure circadian rhythms in humans

Full text

84 How to measure circadian rhythms in humans – Wirz-Justice
MEDICOGRAPHIA, VOL 29, No. 1, 2007
years of research, they laid the basis for the formal
properties of the human circadian system analo-
gous to that developed by Pittendrigh and Daan for
rodents.3
The timing and structure of sleep and waking is
considered to arise from interactions between the
biological clock or circadian pacemaker (designat-
ed “process C”) and a sleep homeostatic process
dependent on duration of prior time awake (“pro-
cess S”) (Figure 1).4This 2-process model is appli-
cable not only to the sleep-wake cycle, but also to
the understanding of the temporal patterns of near-
ly every neuroendocrine, physiological, and psycho-
logical function. The model has proved extremely
useful for understanding a variety of sleep distur-
bances, and can be used to interpret apparent rhyth-
mic abnormalities in depression, as well as provid-
ing a framework for specific therapeutic approaches.
Characteristics of the circadian clock
Biological clocks help us keep time on this rotating
planet. The advantage of an internal clock to regu-
late sleep and wakefulness within the appropriate
The scientific study of human circadian rhythms
began when a curious sleep researcher asked
clever questions and went ahead to test them
on himself. Nathaniel Kleitman spent a month in a
dark underground cave in 1938, having developed
an “apparatus for determining and for recording
motility and rectal temperature during sleep.”1
Kleitman’s monitoring of bed movements by means
of a primitive polygraph to produce a continuous
readout of motor activity anticipated measurement
techniques that have only recently become prac-
tical thanks to advances in microelectronics. He
clearly demonstrated that under constant dim en-
vironmental conditions sleep did not retain its 24-
hour pattern, but shifted later day by day. He also
tried to live on a “28-hour day,” as a test of whether
the usual 24-hour cycle might simply be a reaction
to the outside world. The “28-hour day” is a tech-
nique now used to separate the sleep-wake cycle
(which can more or less follow this long day) from
the endogenous circadian cycle (which cannot).
Two decades later, Jürgen Aschoff and Rütger
Wever created a more comfortable underground
“bunker” for human temporal isolation experiments,
measured motility (by means of sensors in the bed
and floors), rectal temperature, urine output, and
many other physiological and behavioral variables,
and concluded that humans have endogenous cir-
cadian cycles like plants and mice and flies.2They
also placed subjects on days of varying lengths, and
tested to what extent the biological clock could syn-
chronize to the given periodicity. In more than 25
UPDATE
T
he biological clock drives all circadian rhythms in humans,
whether relative to neurobehavioral function, hormones,
physiology, or behavior. The most obvious rhythm is the
sleep-wake cycle, which differs in timing across individuals
(“chronotype”—from early-morning larks to late-night owls).
However, not all changes in sleep-wake cycle behavior are a con-
sequence of abnormal clock function. Knowledge of the formal
properties of the circadian system, the role of zeitgebers for ad-
equate synchronization to the 24-hour day, and how sleep is reg-
ulated, has led to the development of stringent protocols to in-
vestigate the characteristics of circadian rhythms and sleep. These
studies have provided gold standards for estimating circadian
amplitude and phase, and have identified the most useful phys-
iological or hormonal markers. We are now at the second stage
of trying to develop simpler markers for ambulatory use, which
provide a reasonable estimate of circadian phase. Chronobiology
requires long-term measurement over at least one 24-hour cycle,
and new microchip technologies permit noninvasive and contin-
uous data collection over many days and weeks (eg, actimetry).
The next decade of research will surely yield further insights into
human circadian clock function and its pathologies.
Medicographia. 2007;29:84-90. (see French abstract on page 90)
Keywords: human circadian system; sleep regulation; forced
desynchrony; constant routine; ambulatory monitoring;
melatonin

Anna WIRZ-JUSTICE, PhD
Centre for Chronobiology
Psychiatric University Clinics
Basel, SWITZERLAND
HOW TO MEASURE
CIRCADIAN RHYTHMS IN HUMANS
by A. Wirz-Justice, Switzerland
Address for correspondence: Professor Anna Wirz-Justice, Centre for Chronobiology,
Psychiatric University Clinics Basel, Wilhelm Klein Strasse 27, CH-4025 Basel,
Switzerland (e-mail: anna.wirz-justice@unibas.ch)
SELECTED ABBREVIATIONS AND ACRONYMS
DLMO dim light melatonin onset
MCTQ Munich Chronotype Questionnaire
MEQ Morning-Eveningness Questionnaire
SCN suprachiasmatic nuclei

er can entrain, and where endogenous rhythmicity
retains its freerunning period (as originally shown
by Kleitman, Aschoff, and Wever).
It should be noted that this natural periodicity, τ,
is not only a genetically determined characteristic:
τis subject to “after-effects,” ie, is changed by what-
ever environmental light pattern and intensity the
subject was exposed to prior to the study,2such as
phases is that physiology and behavior can antic-
ipate transitions between day and night and not
merely react to them. A circadian clock not only
generates a cycle to match the solar day, it must
also maintain an appropriate phase relation to it.
This process of optimal synchronization with the
environment is called entrainment, and is mediat-
ed by periodic stimuli (“zeitgebers”) acting on the
clock. The endogenous period of the circadian pace-
maker under time-free conditions (as in a cave or
a bunker) is known as τ, and the phase relation be-
tween rhythm and zeitgeber during stable entrain-
ment is defined as ψ(eg, the difference between the
phase of a given circadian rhythm such as sleep
onset and the phase of a zeitgeber such as dusk or
dawn) (Figure 2).2
Individual differences in τmay lead to different
ψ—the best known example is a person with short
τbeing a “lark” chronotype, and someone with long
τbeing an “owl” chronotype.5However, τis not the
only factor that influences phase: sensitivity to light
or zeitgeber strength (when and how long a person
is exposed to what wavelengths and intensities of
light), and amplitude of the circadian pacemaker,
are also determinants.
The most important zeitgeber is light, providing
the photic signal for day and night as well as the
seasons. The master circadian clock in the suprachi-
asmatic nuclei (SCN) consists of two coupled os-
cillatory systems that respond to dawn and dusk.6
The change in daylength with seasons is mimicked
in many species by changes in the duration of ac-
tivity and rest (α:ρ). Three main steps are impor-
tant for biological clock function (Figure 1): input
(zeitgebers, retina) ==> SCN circadian pacemaker
(eg, clock genes, neurotransmitters/peptides) ==>
output (pineal melatonin synthesis, thermoregu-
lation, etc). These factors then interact with the
sleep-wake homeostat to regulate, continuously in
time, sleep propensity and sleep architecture, and
influence phenomena as different as mood and per-
formance or hormonal output.
Which circadian clock characteristics
do we want to measure?
A graphic representation of the various character-
istics of the circadian system is shown in Figure 2.
First, under entrained conditions, the sleep period
remains at a stable phase angle with respect to the
light-dark cycle (ψ), with a given activity-rest ratio
(α:ρ); and, second, in the absence of time cues (zeit-
gebers), the sleep period shifts later and later each
day following the frequency of the endogenous
pacemaker (τ).
Freerunning period (τ)
The freerunning period (τ) is a characteristic of the
circadian pacemaker that can only be measured in
humans using very elaborate protocols: either in a
time-isolation environment in dim light,2whereby
the endogenous rhythmicity reveals itself in a “free
run,” or in a “forced desynchrony” protocol7where
the given sleep-wake cycle is much longer or short-
er than the range to which the circadian pacemak-
UPDATE
85
How to measure circadian rhythms in humans – Wirz-Justice MEDICOGRAPHIA, VOL 29, No. 1, 2007
Circadian
sleep-wake cycle,
neurobehavioral
performance, mood,
cortisol, melatonin,
temperature,
heart rate, etc.
Output
LIGHT 1
4
3
2
Zeitgebers
12 15 18 21 24
Time of da
y
(h)
Phase angle DLMO
to sleep midpoint
Phase angle DLMO
to the LD cycle
DLMO
Phase angle sleep
midpoint to the
LD cycle
SleepDarkLight
Successive days
τ (free run)
ψ (entrained)
α:ρ (activity:rest
duration)
03 06 09 12
αρ
Figure 1. Circadian and homeostatic regulation of sleep. Two major pro-
cesses are involved in driving the circadian sleep-wake cycle as well as
all other behavioral and neuroendocrine outputs: their known anatomical
correlates are schematically represented: (1) retina; (2) suprachiasmatic
nuclei (SCN); (3) hypothalamus: anterior (ventrolateral preoptic nucle-
us—sleep-promoting); posterior (tuberomammillar nucleus —histamine;
orexin [A/B]-producing neurons [wake-promoting]); (4) midbrain and
pons (locus coeruleus [NE]; raphe nuclei [5-HT]; pedonculopontine
tegmentum and laterodorsal tegmentum [ACh]).
Abbreviations: 5-HT, serotonin; ACh, acetylcholine; NE, norepinephrine.
Figure 2. Schematic characteristics of the sleep-wake cycle. The sleep
period (green bars) is plotted on consecutive days with respect to external
time (the light:dark [LD] cycle) and internal time (the circadian phase
marker dim light melatonin onset [DLMO], green circles). See text for
details.

urine) are collected under dim light and controlled
posture conditions, the melatonin rhythm provides
an optimal marker of circadian phase in humans.
“Dim light melatonin onset” (DLMO) is the easiest
marker of body clock time we have, because it can
be feasibly measured in saliva before a person goes
to sleep.12 If an entire 24-hour rhythm is measured,
phase can be defined as required—at the peak, the
midline crossing point, offset of secretion, etc. How-
ever, many studies measure only part of the rhythm,
taking evening samples under controlled conditions
before sleep to measure when melatonin synthesis
begins (DLMO). To understand putative abnormal-
ities, one can measure the phase angle of the sleep
midpoint to external time (in Figure 2, dusk; but
dawn is equally valid), or the phase angle of inter-
nal time as measured by DLMO to the LD cycle or to
sleep midpoint.
However, melatonin assays are not (yet) a rapid
or an easily available method for everyday diagnos-
tic use. A large, carefully controlled study of mela-
tonin rhythms and sleep timing in the same sub-
jects has found a close phase-relationship of this
marker of internal time to the sleep midpoint (Fig-
ure 3).13 An algorithm was developed that is of great
clinical utility, since sleep midpoint can be used as
a reasonable (indirect) estimate of circadian phase
(this is valid only if sleep is not too disturbed14). In
addition, the algorithm allows calculation of when
bright light therapy should be applied with respect
to internal (ie, circadian) and not external (clock)
time. This is the first example of applying circadi-
an principles to determine individual timing of a
treatment: the important therapeutic consequences
are higher remission rates to light therapy than
when prescribing the same clock time for everyone.
We thus can use an everyday, straightforward de-
termination of sleep midpoint (calculated from the
MEQ, MCTQ, or a week of sleep logs) to provide a
rough estimate of an individual’s internal clock time.
There are provocative indications that the same
phase-advancing strategy with early morning light,
known to be beneficial for treatment of winter de-
pression is key to sustained improvement in non-
seasonal depression. Intriguing data from a pre-
liminary study in depressed bipolar patients (on
lithium) treated by sleep deprivation, indicates that
morning light therapy individually timed to max-
imize a circadian rhythm phase advance not only
sustained the rapid sleep deprivation response, but
patients continued to improve (not found with light
given at 11 AM to all) (F. Benedetti, personal com-
munication).
Amplitude
The amplitude of a circadian oscillation is an im-
portant characteristic. When amplitude is low, a
zeitgeber can theoretically elicit larger phase shifts
than when amplitude is high. Measuring amplitude
of clock function is, however, rather difficult in
practice, and there is only indirect evidence from
circadian rhythms of melatonin or temperature that
amplitude can be diminished (by very specific tim-
ing of a light pulse) or augmented (by increasing
light intensity/duration).
the duration of the photoperiod (τis longer in win-
ter than in summer8). Thus, a measured τcould re-
flect behavioral differences with respect to light ex-
posure, rather than just a genetic difference.
Novel techniques now in development can mea-
sure the τof clock gene expression in tissue cultures
from skin biopsies (this yields a τfor “peripheral”
clocks, whose exact relationship to the τexpressed
by the central clock in the SCN is not yet known).9
Phase angle (ψ)
The simplest way to find out about anyone’s pre-
ferred phase position is to ask their preferred sleep
times on free days. Habitual bedtime and wake-up
time on free days are obviously chosen because this
is the comfortable ψfor that individual. The sleep
pattern gives a first simple estimate of a person’s
phase, commonly known as chronotype, which
ranges from extreme early birds to extreme late-
night owls. The most reliable phase marker with
respect to sleep timing is the sleep midpoint (sleep
onset time -- wakeup time / 2). Questionnaires about
sleep preferences have long been used to better
quantify this characteristic. The classic Horne-Ost-
berg Morning-Eveningness Questionnaire (MEQ)10
is now available as an online self-assessment (Au-
toMEQ) with personalized feedback (www.cet.org).
Recently, the Munich Chronotype Questionnaire
(MCTQ) has been developed,5validated against sleep
logs, and also automated (www.imp-muenchen.de);
it is available so far in English, German, Spanish,
Italian, French, and Dutch. Objective measurement
of rest-activity cycle timing can be made with ac-
timetry.11
To go a step further, beyond sleep phase to inter-
nal clock phase, we need a good output of the circa-
dian clock that can be reliably measured. The pineal
hormone melatonin fulfils this role admirably.12 All
species that secrete melatonin do so at night; light
immediately suppresses its synthesis. It has been
well established that if samples (saliva or blood or
UPDATE
86 How to measure circadian rhythms in humans – Wirz-Justice
MEDICOGRAPHIA, VOL 29, No. 1, 2007
80
70
60
50
40
30
20
19 20 23 1
MEQ score
Morning chronotype
Intermediate chronotype
Evening chronotype
DLMO time (h)
21 22 24
Figure 3. Correlation of internal circadian time with chronotype in winter
depression. The timing of dim light melatonin onset (DLMO, threshold
of 3 pg/mL) is strongly correlated with the morningness-eveningness score
on the Morning-Eveningness Questionnaire (MEQ) in patients with winter
depression (r= --0.81, N=35, P<0.001). The wide spread of chronotype is
similar to that found in the general population. Redrawn from unpub-
lished data related to reference 13 with permission from M. Terman.

wakefulness. Thus, actimetry may be used to doc-
ument changes in entrainment state related to ef-
ficacy of a given treatment (whether pharmaceuti-
cal or not). Indeed, a recent study in patients with
seasonal affective disorder showed a delayed rest-
activity cycle and low activity when depressed, with
increased activity and better synchronization fol-
lowing clinical improvement with light therapy.15
There has not yet been very much long-term activ-
ity monitoring in major depression. These patients
may not show a single or consistent abnormality,
but rather, large interindividual differences in their
rest-activity cycle patterns. This is what we are see-
ing in an ongoing study of 6 weeks actimetry in ma-
jor depression during pregnancy (Figure 4). This
unstable rest-activity cycle is indicative of poor cir-
cadian entrainment, and may contribute to some
aspects of the illness.
Over the years, validation of actimetry by sleep
EEG in healthy subjects led to the development of
analysis programs for actimetric “sleep” that pro-
vide an estimate of a number of classic parameters
(sleep onset, wake-up time, wake bouts, sleep effi-
ciency, etc). Analysis programs for circadian vari-
ables (of which there are a variety) can provide an
estimate of, eg, relative amplitude (maximum-min-
imum), intradaily stability (estimates of strength of
coupling to zeitgebers), and interdaily variability
(degree of fragmentation).16
In conclusion, actimetry is an easy, noninvasive,
and relatively inexpensive tool that deserves more
use in the clinic. It provides objective verification of
chronotype (time of going to bed and waking up),
and documents changes in sleep-wake patterns dur-
ing illness and following treatment. A minimum
of 1 week’s recording is recommended to compare
the pattern of work and free days and to reduce vari-
ability.
Why circadian clock characteristics
are not easy to see
Measuring the rest-activity cycle is a first step in
looking at 24-h patterns of behavior. However, to
dig deeper and look at endogenous circadian pace-
maker characteristics requires special techniques
and validated markers. It has long been recognized
that the overtly measured rhythm of a given vari-
able over 24 hours is not only determined by the
biological clock and sleep homeostat, but also by
zeitgebers such as light, or behavior, which can
have direct or indirect effects on many functions,
so-called “masking.”17 Some examples are shown in
Figure 5 (next page).18-20
Masking modifies the pattern of daily rhythms
that are measured in naturalistic environments. As
shown in Figure 5, postural changes rapidly affect
thermoregulation,18 as does sleep,19 thus giving a
false estimate of amplitude and phase when look-
ing at the complete 24-hour measured curves. Light
in the evening can suppress melatonin or delay its
onset, even at lower intensities (>100 lux).20 Thus, in
order to elucidate the characteristics of the endoge-
nous circadian pacemaker, it has been necessary to
develop protocols that control for masking effects.
How can we measure the circadian
rest-activity cycle?
Actimetry is a noninvasive technique for ambula-
tory monitoring of rest-activity (which is not neces-
sarily always congruous with sleep-wake) cycles.11
It is the equivalent of the running wheel for ham-
sters and mice in human circadian biology, with the
same advantages of measurement over longer pe-
riods of time than in sleep research (1 to 2 nights’
polysomnography). Additionally, 24-hour monitor-
ing can reveal unusual patterns of rest and activi-
ty that provide information quite different from the
sleep EEG (eg, timing and duration of daytime naps).
Familiarity with the animal literature on abnormal
rest-activity cycles and the formal properties of the
mammalian circadian system (eg, τ, ψ, see above)3
can help interpret the observed phenomena. How-
ever, it must be clearly recognized that actimetry
does not necessarily reflect the underlying circadi-
an clock characteristics.
A tenet of human chronobiology is that adequate
entrainment means better sleep and higher alert-
ness, and better cognitive state and mood during
UPDATE
87
How to measure circadian rhythms in humans – Wirz-Justice MEDICOGRAPHIA, VOL 29, No. 1, 2007
Figure 4. Long-term activity
monitoring in depression. An
example of an activity moni-
tor, the size of a wristwatch
(Cambridge Neurotechnolo-
gy®), is shown being worn
on the nondominant hand.
Movements are collected at
1- to 2-minute intervals and
stored until readout. The
circadian rest-activity cycle
is double plotted, ie, day 1
and 2 on one line, day 2 and
3 on the next line, etc. This
double plot makes visualiza-
tion of shifted and irregular
rhythms easier. On the ver-
tical axis of each line is the
amount of activity per unit
time: the higher the bar the
greater the activity (blacker).
Upper panel: 6 weeks’ record-
ing of a control pregnant
woman with regular sleep
patterns. Middle and lower
panels: examples of irregular
rest-activity cycles and dis-
turbed nights in pregnant
women suffering from major
depression (Wirz-Justice, un-
published data).
Successive days
Time of day (h)
24 12 1224 24

Forced desynchrony protocol: Under time-free
conditions subjects are asked to sleep on a very
short or very long cycle (eg, a 20-hour or 28-hour
day).21 The circadian system can no longer entrain
to these extremes, and remains at its endogenous
period (usually longer than 24 hours); this means
that over the entire protocol, sleep occurs at every
circadian phase. Post hoc analysis allows a “pure”
circadian and a “pure” sleep homeostatic compo-
nent to be educed from any variable measured (psy-
chomotor vigilance to sleep EEG parameters21 to
subjective mood state (eg, see Figure 3 in refer-
ence 22).
Constant routine protocol: Under time-free con-
ditions, subjects are kept in a constant semirecum-
bent posture in bed, under dim light, controlled
temperature, and humidity, and given small iso-
caloric snacks and water every 1 to 2 hours during
a period of 40 hours of total sleep deprivation.23 This
protocol minimizes masking and provides valid es-
timates of circadian rhythm amplitude and phase
(eg, core body temperature, melatonin).24
Modifications of the above have been developed
that are easier for patients and still reveal circadian
information:
Constant bed rest protocol: Ad libitum sleep al-
lowed instead of sleep deprivation (easier for pa-
tients). Certain circadian rhythms can be measured
under these conditions (eg, sleep propensity, rapid-
eye movement [REM] sleep; see, for example, Fig-
ure 3 in reference 25)
Multiple nap protocol: By scheduling longer naps
over the 24-hour day (eg, 150 minutes awake: 75
minutes asleep) sleep pressure (process S) does not
accumulate and the circadian rhythms of many pa-
rameters thereby usually masked emerge very clear-
ly (eg, subjective sleepiness; see Figure 3 in refer-
ence 26).
The forced desynchrony protocol in particular has
dissected out the relative contributions of the sleep-
wake homeostat and circadian pacemaker to a large
number of neurobehavioral and physiologic func-
tions, which provide the basis for suggested mea-
sures in Table I. These are exhaustive and long pro-
tocols, and thus expensive in terms of recruitment
effort, time in the laboratory, and 24-hour contin-
uous monitoring. Each has its respective advan-
tages, and elegant studies over the last decade have
consolidated the database, established standard val-
ues, and led to a qualitative hierarchy of “which
measures are suitable for which question.” The ma-
jor questions of what changes in ψ, phase relation-
ships between dawn and dusk (α:ρ), or circadian
amplitude occur in major depression require labo-
ratory studies that will help elucidate whether clock
dysfunction is a core feature of the illness.
Markers of the clock
In summary, Table I provides a short list of puta-
tive clock markers for researchers and clinicians.
The list additionally contains external zeitgebers,
since these impact on the structure of sleep and
wake—from the important role of social zeitgebers
to putative conflicting zeitgebers as found in shift
To measure endogenous circadian
rhythms —back to the lab!
When attempting to document circadian rhythm
disturbances, there is always the question of how
much sleep “interferes” with what one measures.
And often, one wants to measure sleep disturbances
as well, because of their obviously important role in
the illness. The 2-process model4is useful here in
selecting the appropriate markers that reflect either
the more circadian or more sleep-related homeo-
static factors. The accepted “gold-standard” proto-
cols are the:
UPDATE
88 How to measure circadian rhythms in humans – Wirz-Justice
MEDICOGRAPHIA, VOL 29, No. 1, 2007
36.6
36.5
36.4
10
N=8
11
CBT (°C)CBT (°C)
Salivary melatonin (pg/mL)
Time of day (h)
Lying down
A
B
C
Standing up
12 13
37.0
36.9
36.8
36.7
36.6
13
N=9
15
14
37.0
36.8
36.6
36.4
36.2 1816
N=36
20
Time of day (h)
Time of day (h)
24 02 04 0622 08
10
8
6
4
2
019
N=7
21 2220 24
23
With sleep
1816
N=7
20 24 02 04 0622 08
Without sleep
Without light
With light
5000 lux
Figure 5. Examples of behavioral and environmental masking effects. Three
examples of masking effects on circadian rhythms are shown: (A) posture and
(B) sleep modify core body temperature (CBT); evening light suppresses mela-
tonin secretion (C).
Redrawn from reference 18 (Panel A):Kräuchi K, Cajochen C, Wirz-Justice A. Thermophysiolog-
ic aspects of the three-process-model of sleepiness regulation. Clin Sports Med. 2005;24:287-300.
Copyright © 2005, Elsevier; reference 19 (Panel B):Kräuchi K, Wirz-Justice A. Circadian clues
to sleep onset mechanisms. Neuropsychopharmacology. 2001;25(5 suppl):S92-S96. Copyright ©
2001, Nature Publishing Group; and reference 20 (Panel C):Wirz-Justice A, Kräuchi K, Cajochen
C, Danilenko KV, Renz C, Weber J. Evening melatonin and bright light administration induce ad-
ditive phase shifts in dim light melatonin onset. J Pineal Res. 2002;36:192-194. Copyright © 2002,
Munksgaard International Publishers.

valid data about the best markers of the clock. To-
gether with the development of new technologies
that allow noninvasive measurement of physiology
and behavior over many days (eg, i-buttons for skin
temperatures,27 actimetry), we have reached the
stage where chronobiology theory meets ambula-
tory research practice. The next decade will surely
bring insights into the complexity of temporal or-
ganization in humans and examples of pathophys-
iology of clock function. 
Many of these approaches are being developed within the
EU 6th Framework Project EUCLOCK (#018741).
workers. Zeitgeber strength (eg, how much light
does an individual receive and at what time of day)
is an important factor for entrainment. Retinal in-
put is a further factor; since light is the major zeit-
geber, eye problems may diminish photic input to
the clock. Blind persons, for whom this signal is ab-
sent, cannot usually synchronize well: they show
phase-delayed or even free-running sleep-wake cy-
cles, indicating that social zeitgebers are not always
sufficient for entrainment.
Thanks to more than a decade of elegant studies
carried out in healthy humans in forced desyn-
chrony and constant routine protocols, we have
UPDATE
89
How to measure circadian rhythms in humans – Wirz-Justice MEDICOGRAPHIA, VOL 29, No. 1, 2007
τ
ψ(chronotype)
ψ
Amplitude
α:ρ
Stability of entrainment
Retinal function
Measured light input
(zeitgeber strength)
Social zeitgebers
Ideal lab protocol
Temporal isolation (free running)
Forced desynchrony
Constant routine (circadian rhythms
of core body temperature,
melatonin, cortisol, heart rate, etc)
Constant routine (circadian rhythm
of core body temperature?)
Actimetry
DLMO over successive days or weeks
Melatonin suppression test
Ambulatory light monitoring
Decline in process S during
sleep
Rise in process S during wake
EEG characteristics
Slow-wave activity (eg, 0.75-4.5 Hz)
in NREM sleep
θ/αactivity in wake EEG
(particularly frontal)
Slow eye movements in wake EEG
Shorter (less perfect) alternatives
DLMO over 3 consecutive weeks
Sleep midpoint (MEQ, MCTQ, sleep
logs, actimetry)
DLMO
Actimetry (relative amplitude
highest-lowest activity)
Sleep logs
Actimetry for at least 7 days
(intradaily stability)
Ophthalmology checkup: (does not
necessarily test circadian photo-
receptor function)
Light logs (time outdoors)
Social Rhythm Metric Questionnaire
Table I. Markers of
circadian rhythms and
the sleep homeostat.
Abbreviations: DLMO,
dim light melatonin onset;
MCTQ, Munich Chrono-
type Questionnaire; MEQ,
Morning-Eveningness
Questionnaire; NREM,
nonrapid eye movement
(sleep).
REFERENCES
1. Kleitman N. Sleep and Wakefulness. Revised and
enlarged edition. Chicago, Ill; The University of Chica-
go Press; 1987.
2. Wever RA. The Circadian System of Man: Results
of Experiments Under Temporal Isolation.New York,
NY: Springer Verlag; 1979.
3. Pittendrigh CS, Daan S. A functional analysis of
circadian pacemakers in nocturnal rodents. J Comp
Physiol (A). 1976;106:233-355.
4. Borbély AA, Achermann P. Sleep homeostasis and
models of sleep regulation. In: Kryger MH, Roth T,
Dement WC, eds. Principles and Practice of Sleep
Medicine. Philadelphia, Pa: WB Saunders Company;
2000:377-390.
5. Roenneberg T, Wirz-Justice A, Merrow M. Life be-
tween clocks: daily temporal patterns of human chro-
notypes. J Biol Rhythms. 2003;18:80-90.
6. Moore RY. Circadian rhythms: basic neurobiology
and clinical applications. Ann Rev Med.1997;48:253-
266.
melatonin profile as a marker for circadian phase po-
sition. J Biol Rhythms. 1999;14:227-237.
13. Terman M, Terman JS. Light therapy for season-
al and nonseasonal depression: efficacy, protocol,
safety, and side effects. CNS Spectr. 2005;10:647-
663.
14. Wright H, Lack L, Bootzin R. Relationship be-
tween dim light melatonin onset and the timing of
sleep in sleep onset insomniacs. Sleep Biol Rhythms.
2006;4:78-80.
15. Winkler D, Pjrek E, Praschak-Rieder N, et al.
Actigraphy in patients with seasonal affective disorder
and healthy control subjects treated with light thera-
py. Biol Psychiatry. 2005;58:331-336.
16. Van Someren EJ. Actigraphic monitoring of move-
ment and rest-activity rhythms in aging, Alzheimer’s
disease, and Parkinson’s disease. IEEE Trans Rehabil
Eng. 1997;5:394-398.
17. Mrosovsky N. Masking: history, definitions, and
measurement. Chronobiol Int. 1999;16:415-429.
7. Dijk DJ, Edgar DM. Circadian and homeostatic
control of wakefulness and sleep. In: Turek FW, Zee
PC, eds. Regulation of Sleep and Circadian Rhythms.
New York, NY: Marcel Dekker, Inc; 1999:111-147.
8. Wirz-Justice A, Wever RA, Aschoff J. Seasonality in
freerunning circadian rhythms in man. Naturwis-
senschaften. 1984;71:316-319.
9. Brown SA, Fleury-Olela F, Nagoshi E, et al. The pe-
riod length of fibroblast circadian gene expression
varies widely among human individuals. PLoS Biol.
2005;3:e338.
10. Horne JA, Östberg O. A self-assessment question-
naire to determine morningness-eveningness in hu-
man circadian rhythms. Int J Chronobiol. 1976;4:
97-110.
11. Ancoli-Israel S, Cole R, Alessi C, Chambers M,
Moorcroft W, Pollak CP. The role of actigraphy in the
study of sleep and circadian rhythms. Sleep. 2003;26:
342-392.
12. Lewy AJ, Cutler NL, Sack RL. The endogenous
SLEEP HOMEOSTAT
CIRCADIAN CHARACTERISTICS

UPDATE
90 How to measure circadian rhythms in humans – Wirz-Justice
MEDICOGRAPHIA, VOL 29, No. 1, 2007

L
’
horloge biologique contrôle tous les rythmes circadiens
chez l’homme, que ce soit pour les fonctions neurocom-
portementales, les sécrétions hormonales, la physiologie
ou le comportement. Le rythme le plus évident est celui du cycle
veille-sommeil dont la synchronisation diffère selon les individus
(« chronotype » – du lève-tôt « au chant du coq », au couche-tard
« au cri du hibou »). Cependant, les altérations du cycle veille-
sommeil ne sont pas tous une conséquence d’un fonctionnement
anormal de l’horloge circadienne. La connaissance des proprié-
tés formelles du système circadien, du rôle des synchroniseurs
(zeitgebers) pour la bonne synchronisation des 24 heures de la
journée et de la façon dont le sommeil est régulé, a permis le dé-
veloppement de protocoles rigoureux pour la recherche des ca-
ractéristiques des rythmes circadiens et du sommeil. Ces études
ont fourni des critères de référence pour l’estimation de l’ampli-
tude et de la phase circadiennes et ont permis l’identification des
marqueurs physiologiques et hormonaux les plus utiles. Nous en
sommes maintenant à la deuxième étape, celle du développement
de marqueurs plus simples à usage ambulatoire pour une esti-
mation raisonnablement exacte de la phase circadienne. La chro-
nobiologie nécessite des mesures à long terme sur au moins un
cycle de 24 heures et les nouvelles technologies à micropuce per-
mettent le recueil continu de données, de façon non invasive et
sur de nombreux jours et semaines (par ex., l’actimétrie). Au
cours des prochaines dix années, la recherche permettra sûrement
une compréhension plus approfondie du fonctionnement de l’hor-
loge circadienne humaine et de sa pathologie.
COMMENT MESURER LES RYTHMES CIRCADIENS CHEZ L’HOMME
18. Kräuchi K, Cajochen C, Wirz-Justice A. Thermo-
physiologic aspects of the three-process-model of
sleepiness regulation. Clin Sports Med. 2005;24:287-
300.
19. Kräuchi K, Wirz-Justice A. Circadian clues to sleep
onset mechanisms. Neuropsychopharmacology. 2001;
25(5 suppl):S92-S96.
20. Wirz-Justice A, Kräuchi K, Cajochen C, Danilenko
KV, Renz C, Weber J. Evening melatonin and bright
light administration induce additive phase shifts in
dim light melatonin onset. J Pineal Res. 2002;36:
192-194.
21. Dijk DJ, Duffy JF, Riel E, Shanahan TL, Czeisler
cillators: why use constant routines? J Biol Rhythms.
2002;17:4-13.
25. Wirz-Justice A, Cajochen C, Nussbaum P. A schizo-
phrenic patient with an arrhythmic circadian rest-
activity cycle. Psychiatry Res. 1997;73:83-90.
26. Cajochen C, Knoblauch V, Kräuchi K, Renz C,
Wirz-Justice A. Dynamics of frontal EEG activity,
sleepiness and body temperature under high and low
sleep pressure. NeuroReport. 2001;12:2277-2281.
27. van Marken Lichtenbelt WD, Daanen HA, Wouters
L, et al. Evaluation of wireless determination of skin
temperature using iButtons. Physiol Behav. 2006;88:
489-497.
CA. Ageing and the circadian and homeostatic regu-
lation of human sleep during forced desynchrony of
rest, melatonin and temperature rhythms. J Physiol.
1999;516:611-627.
22. Boivin DB, Czeisler CA, Dijk DJ, et al. Complex in-
teraction of the sleep-wake cycle and circadian phase
modulates mood in healthy subjects. Arch Gen Psy-
chiatry. 1997;54:145-152.
23. Czeisler CA, Kronauer RE, Allan JS, et al. Bright
light induction of strong (Type 0) resetting of the
human circadian pacemaker. Science. 1989;244:
1328-1333.
24. Duffy JF, Dijk DJ. Getting through to circadian os-