Sleep Medicine 13 (2012) 1071–1078

Contents lists available at SciVerse ScienceDirect

Sleep Medicine
journal homepage: www.elsevier.com/locate/sleep

Original Article

Nocturnal polysomnographic sleep across the menstrual cycle

in premenstrual dysphoric disorder
Ari Shechter a,1, Paul Lespérance b, N.M.K. Ng Ying Kin c, Diane B. Boivin a,⇑
a
Centre for Study and Treatment of Circadian Rhythms, Douglas Mental Health University Institute, Department of Psychiatry, McGill University, Montreal, Quebec, Canada
b
Department of Psychiatry, Centre Hospitalier de l’Université de Montréal, Université de Montréal, Montreal, Quebec, Canada
c
Clinical Research Division, Douglas Mental Health University Institute, Department of Psychiatry, McGill University, Montreal, Quebec, Canada

a r t i c l e i n f o a b s t r a c t

Article history: Objectives: Women with premenstrual dysphoric disorder (PMDD) experience disturbed mood, altered
Received 3 January 2012 melatonin circadian rhythms, and frequent reports of insomnia during the luteal phase (LP) of their men-
Received in revised form 14 May 2012 strual cycle. In this study we aimed to investigate nocturnal polysomnographic (PSG) sleep across the
Accepted 15 May 2012
menstrual cycle in PMDD women and controls.
Available online 30 June 2012
Methods: Seven PMDD women who indicated insomnia during LP, and five controls, spent every third
night throughout a complete menstrual cycle sleeping in the laboratory.
Keywords:
Results: In PMDD and controls progesterone and core body temperature (BTcore) were elevated during LP
Sleep
Menstrual cycle
compared to the follicular phase (FP). Stage 2 sleep showed a significant main effect of menstrual phase
Premenstrual dysphoric disorder and was significantly increased during mid-LP compared to early-FP in both groups. Rapid eye movement
Core body temperature (REM) sleep for both groups was decreased during early-LP compared to early-FP. Slow wave sleep (SWS)
Melatonin was significantly increased, and melatonin significantly decreased, in PMDD women compared to con-
Progesterone trols.
Conclusions: PMDD women who experience insomnia during LP had decreased melatonin secretion and
increased SWS compared to controls. The sleep and melatonin findings in PMDD women may be func-
tionally linked. Results also suggest an altered homeostatic regulation of the sleep–wake cycle in PMDD,
perhaps implicating melatonin in the homeostatic process of sleep–wake regulation.
Ó 2012 Elsevier B.V. All rights reserved.

1. Introduction cause of various methodological issues including differences in

Premenstrual dysphoric disorder (PMDD) is a DSM-IV classified
menstrual cycle-related mood disorder affecting approximately 3–
8% of women [1]. Its occurrence is dependent on, and temporally
defined by, ovarian hormone status across the menstrual cycle.
Typically, symptoms are present during the luteal phase (LP), remit
after menses onset, and are absent throughout the follicular phase
(FP) [2]. Disturbed sleep is among the 11 symptoms that character-
ize the disorder [2], and patients frequently report sleep com-
plaints including insomnia, poor sleep quality, and increased
awakenings during the symptomatic phase [3].
Prior polysomnography (PSG)-based sleep studies focusing on
PMDD have been limited in number and largely inconsistent be-
⇑ Corresponding author. Address: Centre for Study and Treatment of Circadian
Rhythms, Douglas Mental Health University Institute, Department of Psychiatry,
McGill University, 6875 LaSalle Boulevard, Montreal, Quebec, Canada H4H 1R3. Tel.:
+1 (514) 761 6131x2397; fax: +1 (514) 888 4099.
E-mail address: diane.boivin@douglas.mcgill.ca (D.B. Boivin).
1
Current address: New York Obesity Nutrition Research Center, St. Luke’s/
Roosevelt Hospital, New York, NY, USA.
diagnostic criteria, patient group heterogeneity, and limited sam-
pling frequency across the menstrual cycle. Whereas one study
focusing on a group of PMDD women did not show any differences
between controls and patients [4], other studies whose patient
groups included a mix of women with PMDD and severe premen-
strual syndrome (PMS) [5] or premenstrual depression based on
DSM-III criteria [6], observed significantly lengthened rapid eye
movement (REM) sleep onset latency [5], significantly increased
stage 2 sleep [6], or significantly reduced REM sleep [6] in patients
compared to controls regardless of menstrual phase.
The presence of sleep disruption in PMDD, while not a require-
ment for diagnosis, is an important clinical symptom. Insomnia or
hypersomnia is present in approximately 70% of PMDD women [7].
The DSM-IV Research Criteria for Premenstrual Dysphoric Disorder
Checklist used for diagnoses does not differentiate between the
presence of insomnia and hypersomnia [2], however, and, to our
knowledge, no study has reported the relative prevalence of each
within PMDD women. Some of the inconsistencies in the PMDD
sleep literature may be due to studying mixed groups of PMDD wo-
men with insomnia and hypersomnia symptoms and PMDD wo-
men with and without sleep complaints in general. Nevertheless,

1389-9457/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.sleep.2012.05.012
1072 A. Shechter et al. / Sleep Medicine 13 (2012) 1071–1078
sleep deprivation was demonstrated to have an anti-depressive
effect in PMDD [8], supporting the notion that sleep disturbances
are of clinical importance in PMDD. As a growing body of research
indicates that the menstrual cycle has an influence on both subjec-
tive and objective sleep [9], it is increasingly important to docu-
ment sleep in these women by conducting comprehensive
studies across a full menstrual cycle in a homogenous group of
PMDD women with insomnia.
Our aim was to document nocturnal PSG sleep across a full
menstrual cycle in control women and women diagnosed with
PMDD based on DSM-IV criteria who also indicated insomnia
symptoms during the symptomatic phase. Changes in nocturnal
PSG sleep were compared between groups to determine how this
menstrual cycle-related mood disorder affects objective sleep,
and were also considered in light of the altered sex hormone, body
temperature, and nocturnal melatonin secretion profiles associated
with different menstrual cycle phases.
2. Methods
2.1. Participants
Seven PMDD women and five controls were studied. PMDD
diagnoses were based on the Structured Clinical Interview for
DSM-IV (SCID), the Prospective Record of the Impact and Severity
of Menstrual Symptoms (PRISM) [10,11], and a Visual Analogue
Scale (VAS) [10,11] completed daily for at least two menstrual cy-
cles. The reliability and validity of the VAS and PRISM as effective
tools for the prospective tracking of menstrual cycle-related mood
disturbance has been documented [10]. The 11-item VAS (100-mm
bipolar scale, with 0 mm being ‘‘not at all’’ and 100 mm being ‘‘ex-
treme symptoms’’) was based on the DSM-IV criteria for PMDD
diagnoses [2] and included the measures depressed mood, tension,
affective lability, irritability, decreased interest, difficulty concen-
trating, lack of energy, change in appetite, change in sleep patterns,
feeling out of control, and physical symptoms. An individual mean
score for each of the four core PMDD symptoms (depressed mood,
tension, affective lability, irritability) was calculated for days 6–10
after menstruation (FP) and for the last five days of the menstrual
cycle (late-LP). Eligibility criteria required the presence of at least
five of the 11 overall symptoms during late-LP and an increase of
P200% on one, or P100% on two or more of the core symptoms
for the mean late-LP score compared to FP. These diagnostic crite-
ria were developed as prospective criteria to be met, which demon-
strate an unambiguous and clinically relevant worsening of
symptoms during late-LP compared to FP, with the VAS scores
serving to illustrate prospective day-to-day objective differences
in the DSM-IV diagnostic symptoms. The criteria are modeled after,
but are more stringent than, those used by Steinberg et al. [12] and
Steiner et al. [11], which were a 50% worsening in three core symp-
toms or a 100% worsening of one core symptom. All potential
PMDD women met with a psychiatrist (P.L.) twice, at FP and late-
LP, to clinically confirm the PMDD diagnosis. PMDD women with
sleep complaints were recruited. All PMDD women included indi-
cated insomnia symptoms selectively during the premenstrual
phase, with no clinical evidence of subjective sleep disruption
during FP as assessed via clinical consultation. Of the seven PMDD
women, two indicated sleep-onset insomnia, four indicated sleep-
maintenance insomnia, and one reported general insomnia (not
specified). Participants were excluded if diagnosed with another
current Axis I disorder. From the PMDD group, Subjects #2 and
#7 reported one past episode of depression three to four years
prior to study. Seasonal Affective Disorder (SAD) was ruled out
by the Seasonal Pattern Assessment Questionnaire (SPAQ), a tool
used for the assessment of seasonal variations in mood and behav-
ior [13]. PMDD Subject #3 had a global seasonality score of 20 on
the SPAQ, but no diagnosis of SAD was made for this woman based
on the SPAQ and overall clinical assessment. Of interest, she did not
present the typical seasonal variation of sleep with increased dura-
tion during winter. Laboratory visits for PMDD Subjects #1, #2, and
#3 occurred in Winter, PMDD Subject #6 in Spring, PMDD Subject
#4 in Summer and PMDD Subject #5 in Fall. Axis II disorders were
ruled out by clinical evaluation but not systematically with screen-
ing questionnaires. PMDD Subject #1 reported rodent phobia after
a trip to India. PMDD Subject #2 was afraid of airplanes. Age-
matched controls completed the SCID, PRISM, and VAS during
screening and showed no evidence of PMDD or any other psychiat-
ric disorder.
All participants were healthy and drug-free, as confirmed by
medical examination, blood and urine work-up, and toxicology
screening. All had a history of regular menstrual cycles (range:
25–34 ± 3 days), and ovulation was confirmed via a plasma proges-
terone test scheduled on day 21 of the menstrual cycle preceding
experimental procedures. All had no history of gynecological
pathology, were at least six months post-partum, not currently
breast feeding, and free of hormonal contraceptives. At the time
of recruitment, PMDD Subject #4 had a one year old child. Her
insomnia symptoms were present selectively during LP, and were
determined to not be associated with infant sleep patterns. At
the time of recruitment, Control Subject #1 had a seven month
old child but showed no indication of sleep disruption associated
with infant sleep patterns based on screening period actigraphy
and sleep–wake log, as well as no indication of post-partum
depression based on psychiatric evaluations during the screening
phase. Participants had no history of night-shift work or transme-
ridian travel within three months prior to study. Before the exper-
imental month, participants were not allowed to nap during the
day and maintained a regular schedule of 8 h sleep/darkness per
day for at least three weeks confirmed by sleep–wake log, calls
to the laboratory at bed/wake times, and wrist actigraphy for at
least two weeks (Actiwatch, Mini-Mitter, Bend, OR). Chronotype
was assessed with the Horne and Ostberg Morningness–Evening-
ness Questionnaire [14]. Mean morningness–eveningness scores
were comparable between groups and were (mean ± SEM)
50.25 ± 7.73 and 57.57 ± 3.07 for controls and PMDD women,
respectively. This is within the range of ‘‘neither type,’’ i.e., not des-
ignated as morning- or evening chronotypes, though four patients
and one control were ‘‘moderately morning type’’ and one control
was ‘‘definitely evening type’’ [14]. The Douglas Mental Health
University Institute Research Ethics Board approved all procedures
and all participants provided informed consent.
2.2. Experimental design
Participants entered the laboratory for nocturnal PSG sleep
recordings every third night throughout a complete menstrual cy-
cle and left in the morning upon awakening (Fig. 1). The day of the
first laboratory visit ranged from day one to three of the menstrual
cycle (mean ± SD: 1.92 ± 0.51). Participants visited the lab 8–11
times over the course of a menstrual cycle. The first night of sleep
recordings served as an adaptation and diagnostic night to rule out
periodic leg movements in sleep and apnea/hypopnea and was ex-
cluded from subsequent analyses. Wakefulness occurred in regular
room lighting (150 lux). Sleep episodes occurred in complete
darkness (<0.3 lux), lasted 8 h, and were based on the timing of
participants’ habitual sleep/wake schedule maintained during the
screening phase. As part of a larger study, participants underwent
a 24-h period of intensive physiological monitoring under constant
posture (CP) conditions during FP and LP of the menstrual cycle.
Throughout the CP, participants remained in constant conditions,
including a maintained semi-recumbent posture, a time-cue free
 A. Shechter et al. / Sleep Medicine 13 (2012) 1071–1078 1073

Fig. 1. Illustration of laboratory experimental protocol. Participants entered the laboratory for a full night of nocturnal sleep recording once every third night for the duration
of a full menstrual cycle. Polysomnographic sleep and core body temperature (BTcore) were recorded during each of the full sleep episodes, illustrated as black bars. Upon
awakening, a urine sample was collected and assayed for progesterone content. The gray bars illustrate a 24-h period of intensive physiological monitoring under constant
conditions, occurring during the follicular phase and luteal phase, the results of which are the topic of another manuscript. ME, menses; EF, early follicular; MF, mid-follicular;
LF, late follicular; OV, ovulation; EL, early luteal; ML, mid-luteal; LL, late luteal.

environment, iso-caloric snacks served 1x/h, and dim ambient light were also assayed for their content of 6-sulfatoxy-melatonin, the
levels (<10 lux) throughout the waking period before and after the main melatonin metabolite. 6-Sulfatoxy-melatonin concentration
sleep nocturnal sleep episode. Results from the 24-h CP visits will was measured in duplicate using commercially-available radioim-
be reported elsewhere and include all five controls and six of seven munoassay kit (Stockgrand Ltd., Surrey, UK). The sensitivity of the
PMDD women included in the current report. assay was 0.05 ng/ml, with a coefficient of variation ranging from
11.3% to 12.4%.
For all sleep episodes, PSG recordings, including central and
2.3. Measures
occipital electroencephalogram submental electromyogram, and
electrooculogram, were made on a computerized system (Harmonie,
The variation of mood across the menstrual cycle was measured
Stellate Systems, Montreal, QC) at a sampling rate of 250 Hz and
during the screening phase with the 11-item VAS described above.
high- and low-pass filtered at 0.3 Hz and 35 Hz, respectively.
After mean scores were calculated across days 6–10 after menstru-
Periodic limb movements in sleep were screened for by recording
ation and also across the last five days of the menstrual cycle for
electromyograms of the left and right anterior tibialis and scored
each of the four core PMDD symptoms individually (depressed
using established criteria [15]. Respiratory parameters were
mood, tension, affective lability, and irritability), the four core
monitored with bucconasal thermistance and airflow pressure
measures were averaged together to yield a mean score represent-
transducer, and all participants had an apnea/hypopnea index below
ing overall mood symptoms called the VAS-core [11,12].
five. Sleep was visually scored in 30-s epochs according to standard
Core body temperature (BTcore) was continuously monitored
criteria [16]. The amount of each sleep stage was determined and
(4x/min) throughout all sleep episodes via a thermistor (Steri-
expressed as a percent of the sleep period time. Total sleep time
Probe, Cincinatti Sub-Zero Products, Inc., Cincinnati, OH) inserted
was the sum of sleep stages 1–4 plus REM sleep. Sleep efficiency
10 cm into the rectum and connected to an in-house data acquisi-
was total sleep time divided by the time from lights-off to lights-
tion system. Data were inspected visually and by an ad hoc pro-
on multiplied by 100. Sleep onset latency was the time from
gram for probe malfunctions or ‘‘slips,’’ which were discarded. To
lights-off to the first appearance of at least two epochs of stage 1
investigate the variation of nocturnal BTcore across the menstrual
sleep, or the first appearance of any deeper stage. Non-REM sleep
cycle, a sleep-episode average (mean 8-h BTcore, from lights-off to
was the sum of sleep stages 2–4. REM sleep onset latency was the
lights-on) was calculated.
time from sleep onset to the first appearance of REM sleep.
Urine samples were collected each morning upon awakening
and assayed for progesterone concentration. Assays were per-
formed on the Beckman Coulter DxI 800 system using Beckman re- 2.4. Menstrual cycle delineation
agents for chemiluminescence immunoassays (Beckman Coulter
Inc., Brea, CA; coefficient of variation: 6.8%). Morning urine sam- Menstrual cycle length throughout the experimental proto-
ples, to reflect melatonin metabolite excretion during the night, col for PMDD and controls ranged from 24–29 days
1074 A. Shechter et al. / Sleep Medicine 13 (2012) 1071–1078
(mean ± SD: 26.42 ± 1.68 days). Menstrual cycles were normalized
to a standard 28-day cycle based on criteria modeled after Driver
and colleagues [17]. Data obtained during each visit were allocated
into one of eight menstrual phases, including the menses (ME),
early follicular (EF), mid-follicular (MF), late follicular (LF), ovula-
tion (OV), early luteal (EL), mid-luteal (ML), and late luteal (LL)
phases. Menstrual cycle delineation was based on anchor points
at the first day of menstruation, the periovulatory period, as deter-
mined by a surge in luteinizing hormone (LH), and the day of men-
ses for the subsequent menstrual cycle. The LH surge was
determined by a commercially available urine ovulation test
(Church & Dwight Co., Inc., Princeton, NJ), and the periovulatory
period was defined as a three-day window centered on the LH
surge. Days 1–4 after menstruation were assigned to the ME bin.
The three nights before OV were designated as LF, the three nights
occurring before LF were designated as MF, and EF was defined as
the days between ME and MF. The four nights after OV were des-
ignated as EL, the four nights following EL were designated as
ML, and LL was defined as occurring nine nights after OV to the
night before the onset of menses for the subsequent cycle (Tables
S1 and S2). Data from ME were not included in the present results.
2.5. Statistical analyses
Unpaired samples t-tests were used to compare participant
characteristics including age, body mass index, and menstrual cy-
cle length between groups. Two-way ANOVA for repeated mea-
sures (factors: group  menstrual phase) was used to analyze
VAS-core, BTcore, progesterone, 6-sulfatoxy-melatonin, and sleep
parameters across the menstrual cycle. A specific and directional
change (i.e., a reduction) in REM sleep during the luteal phase
was predicted based on our prior findings [9]. This a priori predic-
tion later formed the basis of a series of paired samples t-tests
comparing REM sleep during the follicular and luteal phases (see
Section 3). Effect sizes for between-group differences were calcu-
lated using Cohen’s d, with pooled standard deviation (PMDD
and control). Values of d P 0.8 were considered a large effect size.
3. Results
3.1. Participant characteristics
Controls and PMDD women did not differ in age, body mass
index, or menstrual cycle length (Table 1). Two-way ANOVA
indicated a significant group  menstrual phase interaction for
VAS-core score, with significantly increased late-LP values
observed for PMDD compared to their own FP and to controls
(Table 1). Controls demonstrated no menstrual phase variation
for VAS-core.
Table 1
Participant characteristics.
Characteristic Control
Sample size 5 7
Age, years 30.4 ± 8.20
BMI, kg/cm2 23.5 ± 1.3
Menstrual cycle length, days 26.4 ± 2.3
Mean VAS-core, mm
FPa 3.12 ± 3.87
LLPa 5.57 ± 4.32
PMDD, premenstrual dysphoric disorder; BMI, body mass index; FP, follicular phase; LLP, late luteal phase; VAS-core, mean score of visual analogue scale measures for
depression, tension, affective lability, and irritability; NS, not significant.
Values are mean ± SD.
a
Indicates significant group  menstrual phase interaction for VAS-core, by two-way ANOVA; F1,10 = 37.70, p < 0.001.
b
Indicates significant menstrual phase and group differences for VAS-core, by simple main effects tests; p < 0.001.
3.2. Urinary progesterone
Two-way ANOVA did not show a significant group  menstrual
phase interaction or main effect of group for progesterone (Fig. 2).
A significant main effect of menstrual phase was observed, with
progesterone significantly increased during LP compared to FP in
both groups (Table 2 and Fig. 2).
3.3. 6-Sulfatoxy-melatonin
Two-way ANOVA did not show a significant group  menstrual
phase interaction or main effect of menstrual phase for 6-sulfat-
oxy-melatonin (Fig. 2). A significant main effect of group and large
between-group effect size was observed, with significantly
decreased 6-sulfatoxy-melatonin in PMDD compared to controls
(Table 2 and Fig. 2).
3.4. Nocturnal core body temperature (BTcore)
Two-way ANOVA did not show a significant group  menstrual
phase interaction or main effect of group for BTcore (Fig. 2). A signif-
icant main effect of menstrual phase was observed, with BTcore sig-
nificantly increased during LP compared to FP in both groups
(Table 2 and Fig 2).
3.5. Polysomnographic (PSG) sleep
Two-way ANOVA did not reveal a significant main effect of
menstrual phase or a significant group  menstrual phase interac-
tion for slow wave sleep (SWS). A significant main effect of group
and a large between-group effect size was observed, with signifi-
cantly higher SWS in PMDD compared to controls (Table 2 and
Fig. 2). Two-way ANOVA did not reveal a significant main effect
of group or a significant group  menstrual phase interaction for
stage 2 sleep. A significant main effect of menstrual phase was ob-
served (Table 2). Tukey’s HSD post-hoc comparisons revealed sig-
nificantly increased stage 2 sleep during ML compared to EF
(Fig. 2). Two-way ANOVA did not reveal significant main effects
of group or menstrual phase, or a significant group  menstrual
phase interaction for REM sleep. A paired-samples t-test on pooled
data from controls and PMDD, which was conducted in order to
test the validity of an a priori prediction of reduced REM sleep dur-
ing LP compared to FP [9], confirmed significantly decreased REM
sleep at EL compared to EF (t11 = 2.40, p = 0.04; Fig. 2). Two-way
ANOVA did not reveal a significant main effect of group or a signif-
icant group  menstrual phase interaction for sleep onset latency.
A significant main effect of menstrual phase was observed, though
Tukey’s HSD post-hoc comparisons did not reveal significant differ-
ences between menstrual phases (Table 2 and Fig. 2).
PMDD Control vs. PMDD, p-value
32.0 ± 5.72 p = NS
22.8 ± 2.1 p = NS
26.6 ± 1.5 p = NS
3.06 ± 3.38 p = NS
48.81 ± 16.17b p < 0.001
 A. Shechter et al. / Sleep Medicine 13 (2012) 1071–1078
Fig. 2. The variation of sleep, nocturnal core body temperature (BTcore), urinary
6-sulfatoxy-melatonin (aMt6s), and urinary progesterone across the menstrual
cycle in PMDD women and controls. BTcore values are the mean temperatures
recorded during the nocturnal sleep episode, from lights-out to lights-on. Proges-
terone and aMt6s were assayed from urinary excretion during the sleep period and
collected after awakening. Select sleep measures include sleep efficiency (SE), sleep
onset latency (SOL), rapid eye movement sleep onset latency (ROL), stage 2 sleep
(St. 2), slow wave sleep (SWS), non-rapid eye movement sleep (NREM), and rapid
eye movement sleep (REM). Significant main effects of menstrual phase were
observed for SOL, St. 2, BTcore, and progesterone. REM sleep was reduced during EL
compared to EF. Significant main effects of group were observed for SWS and aMt6s,
with increased SWS and decreased aMt6s in PMDD compared to controls. Values
are mean + SEM. EF, early follicular; MF, mid-follicular; LF, late follicular; OV,
ovulation; EL, early luteal; ML, mid-luteal; LL, late luteal.
1075
Two-way ANOVA did not reveal significant main effects of
group or menstrual phase, or a significant group  menstrual
phase interaction for wakefulness after sleep onset (Table 2).
Two-way ANOVA did not reveal significant main effects of group
or menstrual phase, or a significant group  menstrual phase inter-
action for non-REM sleep (Table 2 and Fig. 2). Two-way ANOVA did
not reveal significant main effects for group or menstrual phase, or
a significant group  menstrual phase interaction for sleep effi-
ciency (Table 2 and Fig. 2). Two-way ANOVA did not reveal signif-
icant main effects of group or menstrual phase, or a significant
group  menstrual phase interaction for REM sleep onset latency
(Table 2 and Fig. 2).
4. Discussion
We examined sleep across the menstrual cycle in PMDD women
with insomnia symptoms during LP and in controls.
Ovulation during the experimental month was confirmed in
participants via increased progesterone during LP compared to
FP. In agreement with others [5,18–22], we found similar proges-
terone levels between groups, which implies that decreased abso-
lute levels of circulating progesterone are unlikely to play a
causative role in the development of PMDD. A positive relationship
was observed between progesterone and BTcore across the men-
strual cycle. Both groups showed significantly increased nocturnal
BTcore throughout LP compared to FP. Progesterone is thermogenic
and likely raises BTcore by inducing an upward shift in the hypotha-
lamic set-point of temperature regulation [23].
Mean nocturnal BTcore was not different between groups. A lim-
ited number of studies have compared nocturnal BTcore between
controls and PMS/PMDD women. In the earliest [24], mean in-
bed rectal temperature was significantly elevated in patients com-
pared to controls throughout EF, LF, ML, and LL. It should be noted
that BTcore recordings in that study were obtained while partici-
pants were at home under ambulatory conditions, and that time-
in-bed was determined from actigraphy and sleep–wake logs.
These experimental conditions contrast with our laboratory-con-
trolled setting in which sleep was polysomnographically docu-
mented. A later study [6] showed no difference between patients
and controls for nocturnal BTcore minimum during sleep episodes
which occurred in the laboratory at EF, LF, EL, and LL. A subsequent
investigation [25] also reported no difference between controls
and PMDD for nocturnal BTcore recorded during in-lab sleep epi-
sodes at MF and LL. Our results indicate that PMDD status does
not alter nocturnal BTcore, and that patients and controls show sim-
ilar variations in BTcore during sleep across menstrual phases.
Elevated nocturnal BTcore during LP in healthy and PMDD wo-
men may contribute to subjective and objective changes in sleep
[9] since sleep propensity [26] and structure [27] vary with circa-
dian phase coincidental with BTcore variations. When evaluating
PSG sleep across the menstrual cycle in healthy and PMDD women,
slight changes were noted for both groups. We observed a men-
strual phase variation for sleep onset latency, though post-hoc
comparisons did not reveal significant differences between men-
strual phases. Upon closer inspection, the decreases in sleep onset
latency at two time points may be due to the scheduling of the two
24-h CP visits which occurred (mean ± SEM) 6.63 ± 0.47 days after
menses (range of days after menses onset: 5–11) and
21.1 ± 0.76 days after menses (range of days after menses onset:
18–25), respectively, in our participants (Fig. 1). Specifically, the
dim lighting (<10 lux), reduced physical activity, constant environ-
mental conditions, and maintained semi-recumbent posture dur-
ing the daytime preceding the nocturnal sleep episode may have
encouraged more rapid sleep initiation at these menstrual phases.
We also observed a significant menstrual phase variation for stage
1076 A. Shechter et al. / Sleep Medicine 13 (2012) 1071–1078

Table 2
ANOVA results.

Variable Main effect of group Main effect of menstrual phase Group  menstrual phase interaction
a
Progesterone, nmol/l p = NS; d = 0.01 p < 0.001 p = NS
aMt6s, ng/ml p = 0.04b; d = 1.01 p = NS p = 0.06c
Tcore, °C p = NS; d = 0.47 p < 0.001d p = NS
SE, % p = NS; d = 0.03 p = NS p = NS
WASO, % p = NS; d = 0.74 p = NS p = NS
SOL, min p = NS; d = 0.50 p = 0.04e p = NS
ROL, min p = NS; d = 0.39 p = NS p = NS
St. 2, % p = NS; d = 0.61 p < 0.04f p = NS
SWS, % p < 0.004g; d = 1.74 p = NS p = NS
NREM, % p = NS; d = 0.22 p = NS p = NS
REM, % p = NS; d = 0.38 p = NS p = NS

aMt6s, 6-sulfatoxy-melatonin; BTcore, core body temperature; SE, sleep efficiency; WASO, wakefulness after sleep onset; SOL, sleep onset latency;
ROL, rapid eye movement sleep onset latency; St.2, stage 2 sleep; SWS, slow wave sleep; NREM, non-rapid eye movement sleep; REM, rapid eye
movement sleep.
p-Values are for two-way between-subjects ANOVA for repeated measures.
d-Values are for between-group Cohen’s d. d P 0.8 was considered a large effect size.
a
F6,60 = 7.43.
b
F6,60 = 2.15.
c
F1,10 = 5.40.
d
F6,60 = 16.65.
e
F6,60 = 2.44.
f
F6,60 = 2.40.
g
F1,10 = 13.35.

2 sleep, which increased during LP compared to FP. This is consis-
tent with the findings of Driver and colleagues [17], who con-
ducted the first systematic study of sleep across the full
menstrual cycle in healthy women. A reduction in REM sleep dur-
ing LP compared to FP has also been reported by us and others,
who observed this variation in healthy women [4,9,17,23,28] and
PMDD patients [4].
The main group difference observed was a significant increase
in SWS in PMDD compared to controls across the menstrual cycle.
This seems to be at the expense of stage 2 sleep, which is decreased
in PMDD, though group-differences were not significant, possibly
due to the small sample size. Previous PSG-based studies compar-
ing PMDD and controls are limited in number and suffer from dif-
ferences in experimental design, menstrual phase of study, and
PMS/PMDD diagnostic criteria, and are therefore inconsistent. A
study of eight women with premenstrual depression and eight
controls examined sleep at EF, LF, EL, and LL and noted significantly
increased stage 2 sleep and decreased REM sleep in patients com-
pared to controls across the menstrual cycle [6]. In a follow-up
study, including 14 women with PMDD and nine controls who
were studied at MF and LL, no significant between-group differ-
ences were observed in sleep parameters [4]. In a group of nine
women with either PMDD or extreme PMS and 12 controls, who
were studied during FP and LL, patients demonstrated significantly
lengthened ROL and a trend for reduced sleep efficiency at both
menstrual phases [5].
The finding of increased SWS in our PMDD women is interesting
to consider in light of the reduced nocturnal melatonin secretion
that characterizes this patient population [18,29]. Indeed, the re-
duced 6-sulfatoxy-melatonin levels observed here in PMDD com-
pared to controls may be functionally related to the alterations in
SWS in these women. Melatonin has soporific properties and can
alter specific aspects of sleep architecture. Exogenous melatonin
administration during the daytime when endogenous levels are
low induced significant reductions in SWS [30] as well as reduc-
tions in stage 3 sleep during the first half of a 16-h sleep opportu-
nity from 16:00–08:00 [31], and during the first 100 min of a
nocturnal sleep episode [32]. Furthermore, exogenous melatonin
also alters the quantitative sleep EEG during a daytime sleep epi-
sode by reducing slow wave activity (SWA) [33]. Authors have
indicated that melatonin may exert its specific effects of reducing
SWS and SWA by acting on GABAA-receptors in a manner reminis-
cent of benzodiazepines [33]. Since slow waves are generated by
interactions between cortical and thalamocortical neuronal net-
works [34], the added presence of melatonin receptors at the tha-
lamic reticular nucleus [35] and the cerebral cortex [36] may
indicate a potential site of action. Interestingly, decreased plasma
GABA concentrations [37] and reduced GABAA-receptor sensitivity
[38] were also reported in PMDD patients compared to controls. It
is possible that in PMDD women the slow wave-generating mech-
anisms in thalamocortical networks may be allowed to function at
a higher baseline level, since the attenuating effects involving the
melatonin or GABA systems may be compromised.
In the current study, PMDD women with insomnia have para-
doxically increased levels of SWS and unaffected sleep efficiency
during nocturnal sleep episodes compared to controls with no
sleep complaints. In another report, women with severe PMS expe-
rienced poorer subjective sleep quality, but no PSG-determined
sleep disturbance, compared to controls [39]. Interestingly, in
those women with severe PMS, poor subjective sleep quality dur-
ing the luteal phase was found to be associated with higher levels
of state-anxiety. This association between mood symptoms and
perceived poor sleep may underlie the dissociation of subjective
and objective sleep measures in these women [39]. Moreover, this
dissociation between subjective and objective sleep characteristics
in PMDD women and controls is reminiscent of comparisons made
between men and women, which determined that, while women
report worsened subjective sleep, PSG-based estimates actually
indicate increased sleep efficiency and fewer awakenings in wo-
men [3]. It is also probable that adequately assessing sleep quality
requires more complex metrics than just quantifying sleep effi-
ciency based on PSG. A 40-h sleep deprivation study suggests that
there are sex-based differences in sleep regulation processes [40].
Specifically, women responded to sleep deprivation with a larger
homeostatic response, as illustrated by significantly increased
SWA power and amplitude during recovery sleep, compared to
men [40]. A further alteration in homeostatic sleep mechanisms
appears to exist in PMDD, which indicates that sleep changes in
these women, like those with major depression, may be the result
of a faulty homeostatic process of sleep regulation [41,42]. As sta-
ted earlier, PMDD women have a perception of poor/insufficient
sleep. However, these women seem to also experience abnormally
 A. Shechter et al. / Sleep Medicine 13 (2012) 1071–1078
high sleep need. PMDD women could therefore experience a ‘‘rel-
ative deficit’’ in sleep, such that despite achieving normal sleep
efficiency, sleep duration, and increased SWS, they may suffer a
relative sleep debt. This is consistent with the observation of con-
comitant increases in SWS and sleep complaints in PMDD women
and implies that an increased sleep opportunity may be useful in
alleviating subjective sleep complaints in PMDD.
The strengths of this investigation include studying all partici-
pants across their complete menstrual cycle. Furthermore, ovula-
tion was confirmed in all participants. Importantly, as opposed to
prior studies, which included a heterogeneous patient group com-
prising both PMS and PMDD, all the women in the PMDD group
met the strict criteria for PMDD diagnoses. All PMDD women in-
cluded in the study reported insomnia symptoms. While we be-
lieve this to be a strength of the study, insomnia in these women
may contribute to the observed changes in SWS and serve as the
basis of the proposed altered sleep pressure and hypothesized ‘‘rel-
ative sleep deficit.’’ As such, these findings would be limited in
their generalizability, and would be applicable to PMDD women
with insomnia symptoms and not necessarily those without sleep
complaints. An interesting avenue of future research may be to
investigate PSG sleep changes in PMDD women with hypersomnia.
Some limitations were present in the current study. Sample size
for both groups is small, thus increasing the potential of a type II
error occurring. Nevertheless, in the intensive study design, each
woman acted as her own control by being studied across a full
menstrual cycle, which we believe adds to the statistical control
of the study. The nature of these laboratory-based experiments
may have affected the results in participants by disrupting their
usual sleeping environment. As both groups might react to this po-
tential stressor differently, we cannot totally exclude the possibil-
ity that environmental factors have contributed to the observed
between-group differences in PSG sleep. Conversely, everyday
stressors and responsibilities could have also been reduced in par-
ticipants because large portions of the menstrual cycle, including
the symptomatic phase in PMDD women, were spent in the labo-
ratory in a controlled environment. Environmental factors could
therefore have influenced mood, as well as subjective and even
objective sleep characteristics. Nevertheless, we believe this poten-
tial bias to be minimal as sleep efficiency was comparable between
groups during both laboratory visits. Along these lines, the con-
stant conditions employed during the CP visits may have created
a relaxing environment, fostering a more rapid sleep onset. Never-
theless, these conditions were applied uniformly for all partici-
pants during both FP and LP, thus limiting its potential as a
contributor to observed group or menstrual phase PSG differences.
Future comprehensive sleep studies based on ambulatory sleep
recordings could clarify these issues.
One PMDD woman included had a high score on the SPAQ,
though a diagnosis of SAD was not made. Moreover, her reported
values did not differ from that of the remainder of the PMDD
group. Finally, some results presented here were interpreted to
indicate a possible disruption in homeostatic sleep regulatory
mechanisms within PMDD women. However, in addition to the
description of baseline SWS included in the current report, to
objectively and precisely document an altered homeostatic sleep
system in these women, it is necessary to examine the dynamics
of SWA decay across the night under baseline conditions and after
increasing homeostatic drive via sleep restriction challenge.
Considering our current findings in light of prior reports under-
scores the preliminary nature of our present SWS results and indi-
cates the need for more refined investigations of quantitative sleep
EEG in PMDD women. Nonetheless, the large effect size observed
for SWS in the present study points toward an altered homeostatic
regulation of sleep in PMDD that might be related to their observed
abnormal melatonin secretion. This would imply that an interac-
1077
tion between homeostatic and circadian processes, which may be
further modulated by the menstrual cycle, plays a role in PMDD-
associated sleep characteristics.
Funding/support
This study was supported by an operating Grant from the Cana-
dian Institutes of Health Research (CIHR). D.B.B. was supported by
scholarships from the Fonds de la recherche en santé du Québec
(FRSQ), CIHR, and Standard Life Foundation.
Disclosure statement
D.B.B. has participated in speaking and advisory engagements
for Servier and Cephalon and is Founder/CEO of Alpha Logik Con-
sultants, Inc. The remaining authors report no potential conflicts
of interest or financial disclosures related to this manuscript.
A.S.’s current affiliation is with the New York Obesity Nutrition Re-
search Center, St. Luke’s/Roosevelt Hospital, New York, NY, USA. He
worked on this project during the course of his graduate studies at
the Centre for Study and Treatment of Circadian Rhythms, Douglas
Mental Health University Institute.
Conflict of interest
The ICMJE Uniform Disclosure Form for Potential Conflicts of
Interest associated with this article can be viewed by clicking on
the following link: doi:10.1016/j.sleep.2012.05.012.
Acknowledgements
We wish to thank the research participants and the staff and
students of the Centre for Study and Treatment of Circadian
Rhythms for their contributions to this investigation. We also
thank Dr. Sylvie Rheaume and Abdelmadjid Azzoug R.N. for medi-
cal supervision, Francine Duquette for dietary information, Véro-
nique Pagé for statistical advice, and Dr. Julie Carrier, Manon
Robert and Sonia Frenette for their advice/assistance with the sleep
analysis, and Zia Choudhry for the sleep recordings.
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.sleep.2012.05.
012.
References
[1] Steiner M. Premenstrual syndrome and premenstrual dysphoric disorder:
guidelines for management. J Psychiatry Neurosci 2000;25(5):459–68.
[2] American Psychiatric Association. Diagnostic and statistical manual of mental
disorders: DSM-IV. 4th ed. Washington, DC; 1994.
[3] Baker FC, O’Brien LM, Armitage R. Sex differences and menstrual-related
changes in sleep and circadian rhythms. In: Kryger MH, Roth T, Dement WC,
editors. Principles and practice of sleep medicine. St. Louis: Elsevier; 2011.
[4] Parry BL, Mostofi N, LeVeau B, Nahum HC, Golshan S, Laughlin GA, et al. Sleep
EEG studies during early and late partial sleep deprivation in premenstrual
dysphoric disorder and normal control subjects. Psychiatry Res
1999;85(2):127–43.
[5] Baker FC, Kahan TL, Trinder J, Colrain IM. Sleep quality and the sleep
electroencephalogram in women with severe premenstrual syndrome. Sleep
2007;30(10):1283–91. Epub 2007/11/01.
[6] Parry BL, Mendelson WB, Duncan WC, Sack DA, Wehr TA. Longitudinal sleep
EEG, temperature, and activity measurements across the menstrual cycle in
patients with premenstrual depression and in age-matched controls.
Psychiatry Res 1989;30(3):285.
[7] Hurt SW, Schnurr PP, Severino SK, Freeman EW, Gise LH, Rivera-Tovar A, et al.
Late luteal phase dysphoric disorder in 670 women evaluated for premenstrual
complaints. Am J Psychiatry 1992;149(4):525–30. Epub 1992/04/01.
[8] Parry BL, Wehr TA. Therapeutic effect of sleep deprivation in patients with
premenstrual syndrome. Am J Psychiatry 1987;144(6):808–10. Epub 1987/
06/01.
1078 A. Shechter et al. / Sleep Medicine 13 (2012) 1071–1078
[9] Shechter A, Varin F, Boivin DB. Circadian variation of sleep during the follicular
and luteal phases of the menstrual cycle. Sleep 2010;33(5):647–56.
[10] Steiner M, Steinberg S, Stewart D, Carter D, Berger C, Reid R, et al. Fluoxetine
in the treatment of premenstrual dysphoria. N Engl J Med 1995;332(23):
1529–34.
[11] Steiner M, Streiner DL, Steinberg S, Stewart D, Carter D, Berger C, et al. The
measurement of premenstrual mood symptoms. J Affect Disord
1999;53(3):269–73.
[12] Steinberg S, Annable L, Young SN, Liyanage N. A placebo-controlled clinical
trial of L-tryptophan in premenstrual dysphoria. Biol Psychiatry
1999;45(3):313–20.
[13] Hardin TA, Wehr TA, Brewerton T, Kasper S, Berrettini W, Rabkin J, et al.
Evaluation of seasonality in six clinical populations and two normal
populations. J Psychiatr Res 1991;25(3):75–87. Epub 1991/01/01.
[14] Horne JA, Ostberg O. A self-assessment questionnaire to determine
morningness–eveningness in human circadian rhythms. Int J Chronobiol
1976;4(2):97–110.
[15] Coleman RM. Periodic movements in sleep (nocturnal myoclonus) and restless
legs syndrome. In: Guilleminault C, editor. Sleeping and waking disorders:
indications and techniques. Menlo-Park: Addison-Wesley; 1982. p. 265–95.
[16] Rechtschaffen A, Kales A. A manual of standardized terminology, techniques
and scoring systems for sleep stages of human subjects. Los Angeles: Brain
Information Service, Brain Research Institute, UCLA; 1968.
[17] Driver HS, Dijk DJ, Werth E, Biedermann K, Borbely AA. Sleep and the sleep
electroencephalogram across the menstrual cycle in young healthy women. J
Clin Endocrinol Metab 1996;81(2):728–35.
[18] Parry BL, Berga SL, Kripke DF, Klauber MR, Laughlin GA, Yen SS, et al. Altered
waveform of plasma nocturnal melatonin secretion in premenstrual
depression. Arch Gen Psychiatry 1990;47(12):1139–46.
[19] Parry BL, Hauger R, LeVeau B, Mostofi N, Cover H, Clopton P, et al. Circadian
rhythms of prolactin and thyroid-stimulating hormone during the menstrual
cycle and early versus late sleep deprivation in premenstrual dysphoric
disorder. Psychiatry Res 1996;62(2):147–60.
[20] Parry BL, Meliska CJ, Sorenson DL, Lopez A, Martinez LF, Hauger RL, et al.
Increased sensitivity to light-induced melatonin suppression in premenstrual
dysphoric disorder. Chronobiol Int 2010;27(7):1438–53. Epub 2010/08/28.
[21] Rapkin AJ, Morgan M, Goldman L, Brann DW, Simone D, Mahesh VB.
Progesterone metabolite allopregnanolone in women with premenstrual
syndrome. Obstet Gynecol 1997;90(5):709–14.
[22] Schmidt PJ, Purdy RH, Moore Jr PH, Paul SM, Rubinow DR. Circulating levels of
anxiolytic steroids in the luteal phase in women with premenstrual syndrome
and in control subjects. J Clin Endocrinol Metab 1994;79(5):1256–60. Epub
1994/11/01.
[23] Baker FC, Driver HS, Paiker J, Rogers GG, Mitchell D. Acetaminophen does not
affect 24-h body temperature or sleep in the luteal phase of the menstrual
cycle. J Appl Physiol 2002;92(4):1684–91.
[24] Severino SK, Wagner DR, Moline ML, Hurt SW, Pollak CP, Zendell S. High
nocturnal body temperature in premenstrual syndrome and late luteal phase
dysphoric disorder. Am J Psychiatry 1991;148(10):1329–35.
[25] Parry BL, LeVeau B, Mostofi N, Naham HC, Loving R, Clopton P, et al.
Temperature circadian rhythms during the menstrual cycle and sleep
deprivation in premenstrual dysphoric disorder and normal comparison
subjects. J Biol Rhythms 1997;12(1):34–46.
[26] Lack LC, Lushington K. The rhythms of human sleep propensity and core body
temperature. J Sleep Res 1996;5(1):1–11.
[27] Dijk DJ, Czeisler CA. Contribution of the circadian pacemaker and the sleep
homeostat to sleep propensity, sleep structure, electroencephalographic slow
waves, and sleep spindle activity in humans. J Neurosci 1995;15(5 Pt.
1):3526–38.
[28] Baker FC, Driver HS, Rogers GG, Paiker J, Mitchell D. High nocturnal body
temperatures and disturbed sleep in women with primary dysmenorrhea. Am
J Physiol 1999;277(6 Pt 1):E1013–21.
[29] Parry BL, Berga SL, Mostofi N, Klauber MR, Resnick A. Plasma melatonin
circadian rhythms during the menstrual cycle and after light therapy in
premenstrual dysphoric disorder and normal control subjects. J Biol Rhythms
1997;12(1):47–64.
[30] Hughes RJ, Badia P. Sleep-promoting and hypothermic effects of daytime
melatonin administration in humans. Sleep 1997;20(2):124–31.
[31] Rajaratnam SM, Middleton B, Stone BM, Arendt J, Dijk DJ. Melatonin advances
the circadian timing of EEG sleep and directly facilitates sleep without altering
its duration in extended sleep opportunities in humans. J Physiol 2004;561(Pt.
1):339–51.
[32] Stone BM, Turner C, Mills SL, Nicholson AN. Hypnotic activity of melatonin.
Sleep 2000;23(5):663–9.
[33] Aeschbach D, Lockyer BJ, Dijk DJ, Lockley SW, Nuwayser ES, Nichols LD, et al.
Use of transdermal melatonin delivery to improve sleep maintenance during
daytime. Clin Pharmacol Ther 2009;86(4):378–82.
[34] Steriade M, McCormick DA, Sejnowski TJ. Thalamocortical oscillations in the
sleeping and aroused brain. Science 1993;262(5134):679–85.
[35] Lindroos OF, Leinonen LM, Laakso ML. Melatonin binding to the anteroventral
and anterodorsal thalamic nuclei in the rat. Neurosci Lett 1992;143(1–
2):219–22. Epub 1992/08/31.
[36] Mazzucchelli C, Pannacci M, Nonno R, Lucini V, Fraschini F, Stankov BM. The
melatonin receptor in the human brain: cloning experiments and distribution
studies. Brain Res Mol Brain Res 1996;39(1–2):117–26. Epub 1996/07/01.
[37] Halbreich U, Petty F, Yonkers K, Kramer GL, Rush AJ, Bibi KW. Low plasma
gamma-aminobutyric acid levels during the late luteal phase of women with
premenstrual dysphoric disorder. Am J Psychiatry 1996;153(5):718–20.
[38] Sundstrom I, Backstrom T. Patients with premenstrual syndrome have
decreased saccadic eye velocity compared to control subjects. Biol Psychiatry
1998;44(8):755–64.
[39] Baker FC, Sassoon SA, Kahan T, Palaniappan L, Nicholas CL, Trinder J, et al.
Perceived poor sleep quality in the absence of polysomnographic sleep
disturbance in women with severe premenstrual syndrome. J Sleep Res
2012. Epub 2012/03/16.
[40] Armitage R, Smith C, Thompson S, Hoffmann R. Sex differences in slow-wave
activity in response to sleep deprivation. Sleep Res Online 2001;4(1):33–41.
[41] Boivin DB. Influence of sleep-wake and circadian rhythm disturbances in
psychiatric disorders. J Psychiatry Neurosci 2000;25(5):446–58.
[42] Borbely AA, Wirz-Justice A. Sleep, sleep deprivation and depression. A
hypothesis derived from a model of sleep regulation. Hum Neurobiol
1982;1(3):205–10.