Psychoneuroendocrinology (2010) 35, 47—55

a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m

j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / p s y n e u e n

Neural mechanisms underlying changes

in stress-sensitivity across the menstrual cycle
Lindsey Ossewaarde a,*, Erno J. Hermans a,b, Guido A. van Wingen a,b,
Sabine C. Kooijman a, Inga-Maj Johansson c,
Torbjörn Bäckström c, Guillén Fernández a,b

a
Radboud University Nijmegen, Donders Institute for Brain, Cognition and Behaviour, Nijmegen, The Netherlands
b
Radboud University Nijmegen Medical Centre, Department of Neurology, Nijmegen, The Netherlands
c
Umeå Neurosteroid Research Center, Department of Clinical Science, Obstetrics and Gynecology, Umeå, Sweden

Received 3 July 2009; received in revised form 19 August 2009; accepted 20 August 2009

KEYWORDS Summary Hormonal fluctuations across the menstrual cycle are thought to play a central role
Menstrual cycle; in premenstrual mood symptoms. In agreement, fluctuations in gonadal hormone levels affect brain
Stress; processes in regions involved in emotion regulation. Recent findings, however, implicate psycho-
Amygdala; logical stress as a potential mediating factor and thus, we investigated whether effects of moderate
Medial prefrontal cortex; psychological stress on relevant brain regions interact with menstrual cycle phase. Twenty-eight
Allopregnanolone; healthy women were tested in a crossover design with menstrual cycle phase (late luteal versus late
fMRI follicular) and stress (stress induction versus control) as within-subject factors. After stress
induction (or control), we probed neural responses to facial expressions using fMRI. During the
late luteal phase, negative affect was highest and the stress-induced increase in heart rate was
mildly augmented. fMRI data of the control condition replicate previous findings of elevated
amygdala and medial prefrontal cortex responses when comparing the late luteal with the late
follicular phase. Importantly, stress induction had opposite effects in the two cycle phases, with
unexpected lower response magnitudes in the late luteal phase. Moreover, the larger the increase in
allopregnanolone concentration across the menstrual cycle was, the smaller the amygdala and
medial prefrontal cortex responses were after stress induction in the late luteal phase. Our findings
show that moderate psychological stress influences menstrual cycle effects on activity in the
emotion regulation circuitry. These results provide potential insights into how fluctuations in
allopregnanolone that naturally occur during the menstrual cycle may change stress vulnerability.
# 2009 Elsevier Ltd. All rights reserved.

* Corresponding author at: Radboud University Nijmegen, Donders 1. Introduction

Institute for Brain, Cognition and Behaviour, Centre for Cognitive
Neuroimaging, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands.
Tel.: +31 024 3666274; fax: +31 024 3610989.
E-mail address: lindsey.ossewaarde@donders.ru.nl
(L. Ossewaarde).
Several psychiatric disorders are associated with periods of
instability in the hormonal milieu across the woman’s repro-
ductive cycle (Steiner et al., 2003). For example, women

0306-4530/$ — see front matter # 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.psyneuen.2009.08.011
48
with premenstrual dysphoric disorder (PMDD) suffer from
symptoms of mood dysregulation only during the late luteal
phase of the menstrual cycle, when progesterone and estra-
diol levels are high and declining (Bäckström et al., 1983). It
has been suggested that most of these premenstrual com-
plaints reported by women with PMDD are related to heigh-
tened stress-sensitivity (Epperson et al., 2007; Kask et al.,
2008). Similar findings have been reported in healthy women.
Studies measuring cognitive performance, cardiovascular
reactivity, and secretion of stress hormones like cortisol
and (nor)epinephrine in response to physical and various
psychological stressors suggest that stress-sensitivity
increases in the luteal phase as compared to the follicular
phase (Marinari et al., 1976; Tersman et al., 1991; Kumari and
Corr, 1998; Moldovanova et al., 2008). These findings suggest
that menstrual cycle-dependent decreases in stress resili-
ence may prime the development of negative mood symp-
toms in the late luteal phase and other periods with high or
declining levels of gonadal hormones. However, the neural
interactions that give rise to such menstrual cycle-dependent
changes in stress-sensitivity are currently unknown.
Gonadal hormones may modulate stress-sensitivity by
affecting neural activity in the amygdala and the medial
prefrontal cortex (mPFC) (Shansky et al., 2004; Herman
et al., 2005). These brain regions play a major role in
modulating the sympatho-adrenomedullary (nor)adrenergic
and hypothalamic—pituitary—adrenal (HPA) axis responses to
stress, and are involved in regulating emotion (Gold et al.,
2002; Phelps and LeDoux, 2005). Indeed, neuroimaging stu-
dies have shown that neural responses in these regions to
emotional stimuli change across the menstrual cycle (Proto-
popescu et al., 2005; Derntl et al., 2008). One of the candi-
date mechanisms that have been proposed for such effects is
through actions of progesterone and its metabolites, most
notably the neuroactive steroid allopregnanolone. Such neu-
rosteroids are potent allosteric modulators of the g-amino-
butyric acid type A (GABA-A) receptor (Belelli and Lambert,
2005). Consistent with this notion, pharmacological studies
have shown that exogenous manipulation of progesterone
affects neural activity in regions similar to those where cycle-
dependent changes were observed (van Wingen et al., 2007,
2008). Therefore, these brain regions may also mediate, via a
GABAergic interaction, the changes in stress-sensitivity
observed across the menstrual cycle.
In the present study, we therefore investigated whether
the menstrual cycle modulates the effects of psychological
stress on functioning of brain areas involved in emotion
regulation. Neural responses in these regions during the late
luteal phase, when women with PMDD are most symptomatic,
were compared with those in the late follicular phase, when
these women are least symptomatic (Sveindottir and Bäck-
ström, 2000). Strongly aversive movie clips were presented
to induce psychological stress (Qin et al., 2009). To assess the
effects of stress induction on the sympathetic nervous system
(SNS) and HPA axis, heart rate was recorded throughout
scanning, and salivary cortisol samples were collected at
baseline and at various time delays. Subsequently, respon-
siveness in emotion regulation circuits was probed by pre-
senting dynamic facial expressions, which robustly elicit
neural activity in these regions (Sato et al., 2004). We
anticipated interactive effects of menstrual cycle phase
and psychological stress in the aforementioned brain regions.
L. Ossewaarde et al.
2. Materials and methods
2.1. Participants
Twenty-eight healthy right-handed women participated in
this study (mean age: 22.8 years, range: 18—38 years). None
had used hormonal contraceptives for the previous three
months and all had regular menstrual cycles (26—32 days
over the past year). Participants had no history of psychiatric
disorders (as determined with the Mini International Neurop-
sychiatric Interview, M.I.N.I) (Sheehan et al., 1998) and were
additionally screened specifically for Premenstrual Dysphoric
Disorder or Premenstrual Syndrome using the Dutch version of
the Moos Menstrual Distress Questionnaire (MDQ) (Moos,
1968). Furthermore, they were all physically healthy and
free of medication, and reported no MRI contraindications.
Data of two women were excluded from data analysis due to
excessive head movement during scanning. The study was
approved by the local ethical review board (CMO Region
Arnhem-Nijmegen, The Netherlands) and all women provided
written informed consent.
2.2. Design and procedure
Women were tested in a crossover design with the counter-
balanced factors menstrual cycle phase (late follicular versus
late luteal; see also Fernández et al. (2003), Weis et al.
(2008)) and stress induction (stress versus control condition)
as within subject factors. To ensure relatively stable and low
levels of endogenous cortisol, the experiment was carried out
after 1400 h. All women were scanned four times on two days
during the menstrual cycle, twice in the late follicular phase
(8—12 days after the onset of menses) and twice in the late
luteal phase (10—14 days after ovulation). For menstrual
cycle phase verification we obtained prospective self-reports
about when their menstruation started, ovulation assess-
ments (as determined by using commercially available ovula-
tion predictor kits estimating the LH peak in urine) and
allopregnanolone concentrations. On each testing day, the
second session started after a break of about 100 min.
2.3. Stress induction
To induce a stressful state, highly aversive movie clips were
shown immediately before the actual task (Henckens et al.,
2009; Qin et al., 2009; van Marle et al., 2009). These clips
consisted of scenes of a movie [Irréversible (2002), Gaspar
Noé] containing maximally aggressive behavior and violence
against men and women. For the control condition, neutral
scenes of another movie were shown [Comment j’ai tué mon
père (2001), Anne Fontaine]. The stressful and the neutral
movie clips were similar in the amount of speech, human
(face) presence, luminance, and language.
2.4. Dynamic facial expression task (DFET)
Immediately after showing either the stressful or neutral
movie clip, women passively viewed blocks of faces morphing
dynamically from a neutral expression into either an angry,
happy, or fearful expression. Grayscale photographs of 10
individuals’ faces were taken from standardized sets (Ekman
Menstrual cycle and stress
and Friesen, 1976; Lundqvist et al., 1998). For each face and
each emotion short computer animated clips were created
using WinMorph 3.01. Each task run lasted 11.5 min and
consisted of 27 blocks of 25 s each: 6 blocks for each emotion
and 9 blocks of low level fixation cross baseline. Each emotion
block contained 50 morphing sequences presented at a rate
of 2 Hz, avoiding repetitions of the same morphing sequence.
Sequences started with a 55% emotional expression and
ended with a 100% emotional expression with two intermedi-
ate steps of 70% and 85% in between. Each frame was
presented for 33 ms. Thus, durations of sequences and inter-
sequence intervals were 133 ms and 367 ms, respectively.
Emotion blocks were presented in a mirrored design avoiding
covariation with linear drift, and adjacent blocks of the same
emotion or fixation cross were avoided. The order of blocks
was counterbalanced across women. Women were instructed
to watch attentively and make a right index finger response
on a button box each time when a fixation block started. This
response was recorded to control for attention.
2.5. MR data acquisition
MRI scans were collected using a Siemens (Erlangen, Ger-
many) TIM Trio 3.0 Tesla MRI scanner equipped with an eight
channel phased array head coil. The following scans were
obtained: Four runs of 363 T2 weighted BOLD images each
(gradient echo EPI, TE/TR: 25/1890 ms, flip angle 808, FOV:
212 mm 212 mm, matrix 64  64, 3 mm slice thickness,
0.3 mm slice gap, 37 ascending slices. To reduce signal
drop-out and geometric distortions, we used a short TE, an
oblique axial angulation (de Zwart et al., 2006), and reduced
echo-train length by means of factor 2 accelerated GRAPPA
(Griswold et al., 2002). Structural scans were obtained using
an MP-RAGE sequence (TE/TR: 2.96/2300 ms, flip angle: 88,
FOV: 256 mm 256 mm 192 mm, voxel size: 1 mm isotro-
pic, GRAPPA acceleration factor 2).
2.6. fMRI data analysis
Data were analyzed using Statistical Parametric Mapping
software (SPM5; Wellcome Department of Imaging Neu-
roscience, London). To allow for T1 equilibration, the first
five EPI-volumes of each run were discarded. The remaining
images were realigned to the first volume, coregistered
to the structural MR image, spatially normalized the stan-
dard Montreal Neurological Institute (MNI) 152 coordinate
space, resampled into 2 mm 2 mm  2 mm voxels, and
smoothed with an isotropic 8-mm full-width half maximum
Gaussian kernel.
Statistical analysis was performed within the framework
of the general linear model. For each of the four runs, the
experimental conditions (angry, fear, and happy) were
modeled as separate box-car regressors, which were subse-
quently convolved with the canonical hemodynamic response
function implemented in SPM5. The six parameters corre-
sponding to the movement parameters obtained from the
realignment procedure were also included in the model, as
well as a high-pass filter with a 1/128 Hz cut-off frequency.
We applied proportional global signal scaling to reduce
effects due to global signal variations between scan sessions.
The single subject parameter estimates of each session and
49
condition obtained from the first level analysis were included
in subsequent second level analyses treating subjects as
random variable. A repeated-measures ANOVA was used, with
menstrual cycle phase (late follicular versus late luteal),
stress (stress induction versus neutral), and emotional
expression (angry versus fear versus happy) as within-subject
factors, with non-sphericity corrections for repeated mea-
sures. Subsequent correlational analyses were performed on
relevant contrast estimate images using the difference in
allopregnanolone level between the late follicular and late
luteal phase as a covariate of interest. Statistical thresholds
were set at P < 0.05, family wise error (FWE) corrected using
Gaussian Random Field theory (Worsley et al., 1996). A small
volume correction (SVC) procedure for a reduced search
region was applied for our a priori regions of interest, namely
the amygdala and mPFC. Specifically, for the amygdala, we
applied a spherical search region with a radius of 10 mm
placed on the center of the amygdala with the MNI coordi-
nates 20, 6, 16 (x, y, z), as determined based on the
Talairach Atlas (Talairach and Tournoux, 1988). Medial pre-
frontal regions were defined as spheres with 20 mm radii
around previously reported peaks of menstrual cycle effects
in this region (MNI coordinates 3, 51, 18 and 15, 57, 15
(x, y, z) (Protopopescu et al., 2005).
2.7. Subjective and physiological measures of
stress
Changes in affect were assessed using the Positive And
Negative Affect Scales (PANAS) (Watson et al., 1988) at three
time points during each test day: at baseline (15 min after
arrival), immediately after the first DFET (75 min after arri-
val), and immediately after the last DFET (180 min after
arrival) coinciding with collection of saliva samples. To
monitor the HPA-axis response, cortisol samples were col-
lected using salivette collection devices (Sarstedt, Rommels-
dorf, Germany). Women were requested not to brush or floss
their teeth or to eat or drink anything but water for 2 h prior
saliva sampling. Salivary sampling consisted of two baseline
measurements (before MRI scanning) and five additional ones
(right after the first DFET, 20 min after the DFET which is
during the break, right before the last movie, right after the
last DFET and 15 min after leaving scanner). All samples were
stored at 20 8C until analysis. Samples were prepared for
biochemical analysis by centrifuging at 3000 rpm for 5 min,
which resulted in a clear supernatant of low viscosity. Sali-
vary-free cortisol concentrations were determined employ-
ing a chemiluminescence assay (CLIA) with high sensitivity of
0.16 ng/ml (IBL, Hamburg, Germany). To assess the sympa-
thetic nervous system response, heart rate (HR) was recorded
continuously during exposure to the aversive movie clips and
the neutral control clips by means of an infrared pulse
oximeter attached to the left index finger. To analyze the
heart rate signal, movement-related and other measurement
artifacts were rejected using in-house software. Heart rate
data of four women were discarded because of excessive
movement artifacts in the recorded signal. A commercial MR
compatible eye-tracking device (MEyeTrack-LR, Senso Moto-
ric Instruments, Teltow, Germany) was used to monitor
attentive viewing of the movie scenes and the DFET through-
out the scan procedure. Reaction time data recorded during
the DFET (response latency to the onset of fixation baseline
50
blocks) were filtered using a 3 standard deviation outlier
removal cut-off.
2.8. Gonadal hormone sampling and analysis
In order to measure gonadal hormone levels including allo-
pregnanolone, 10 ml of saliva was collected at each test day
(5 min after arrival). Saliva was collected in plastic tubes and
kept frozen at 20 8C. At analysis the thawed samples were
centrifuged 15 min 1000  g at 4 8C and the supernatant used
for analysis. Saliva collected in the follicular phase, 3.5—
12 ml, was analyzed as one sample; while the luteal phase
collected saliva was divided and analyzed in duplicate (2.2—
7.1 ml/sample). With larger volumes samples were divided
into several tubes during the diethyl ether extraction, with a
maximum of 3 ml saliva per extraction with 5 ml diethyl
ether. Given the experimental nature and novelty of the
allopregnanolone assessment based on saliva the entire
samples were used for this analysis. Yet, it has been shown
that progesterone and allopregnanolone levels are closely
and positively correlated with each other (Wang et al., 1996).
2.8.1. High performance liquid chromatography
(HPLC)
The HPLC system consisted of a Waters 1515 Isocratic Pump,
delivering the mobile phase (methanol: water, 60:40, v/v) at
a flow rate of 1.0 ml/min and a Symmetry C18 3.5 mm
4.6 mm  75 mm separation column (Waters), heated to
45 8C. Detection of retention times of standards and cross-
reacting steroids was validated with spiked samples of saliva
at 206 nm using a Waters 2487 Dual l Absorbance Detector.
The detector output was recorded on a PC-based Waters
Breeze Chromatography Software (version 3.20). In the pre-
parative HPLC 3 ml fractions were symmetrically collected
around the retention time for allopregnanolone for further
analysis with RIA. No cross-reacting steroids had retention
times close to the collected fraction (Turkmen et al., 2004).
2.8.2. Radioimmunoassay (RIA)
The allopregnanolone concentration was determined using
an allopregnanolone antiserum commonly used and kindly
provided by Dr RH Purdy, Department of Physiology and
Pharmacology, the Scripps Research Institute, La Jolla, CA,
USA (Purdy et al., 1990). The antiserum is raised against
3a-hydroxy-20-oxo-5a-pregnan-11-yl carboxymethyl ether
coupled to bovine serum albumin, and used in a dilution of
1/5000. The standard curve (range from 0 to 400 pg) and
extracted samples were dried to dryness under nitrogen gas
before antiserum solution was added. The antibody solu-
tion was prepared using antiserum, [11,12]3H-allopregna-
nolone (Perkin Elmer Life Sciences, Boston, USA), 65 mM
boric acid (Merck), bovine serum albumin 100 mg/ml
(Sigma, St. Louis, USA), and human gamma globulin solu-
tion 20 mg/ml (Octapharma, Sweden). The solution was
allowed to equilibrate overnight at 8 8C. Antibody solution
(200 ml) was added to all standard and sample tubes incu-
bated overnight at 8 8C. The proteins were precipitated by
adding saturated ammonium sulfate, mixed and centri-
fuged. The supernatant was counter in a (RackBeta Wallac,
Finland) scintillation counter. The sensitivity of the assays
was 19 pg. The recovery of allopregnanolone was calcu-
L. Ossewaarde et al.
lated by using spiked saliva samples and was mean 76.2%
range (62—92%) and the results are compensated for recov-
ery. The intraassay coefficient of variation was 21% in these
saliva samples.
3. Results
3.1. Behavioral and physiological results
3.1.1. Hormone assessment
The saliva concentrations of allopregnanolone were higher
in the late luteal phase (mean  SD, 0.050  0.016 nmol/l)
than the late follicular phase (mean  SD, 0.035 
0.010 nmol/l) (t25 = 4.67, P < 0.001). Kolmogorov—Smirnov
tests confirmed that the distributions of allopregnanolone
concentrations in the two cycle phases did not deviate
from a normal distribution (both Z < 1). Moreover, distri-
butions contained no outliers. Note that levels of allopreg-
nanolone are tightly linked to levels of progesterone
(Wang et al., 1996). Although normally estradiol and/or
progesterone levels are used for confirmation of menstrual
cycle phase, these data therefore do indicate that all
participants were tested during the intended menstrual
cycle phase.
Salivary cortisol was used to assess the HPA-axis stress
response. A 2 (cycle phase) by 5 (time: five post-baseline
time points) ANOVA revealed a significant downward pattern
in cortisol for both cycle phases over time (F(4,20) = 12.546,
P < 0.01). There were neither main effects of stress induc-
tion or menstrual cycle phase, nor interaction effects (all
P > 0.05), most likely due to the short delay between the
two sessions per cycle day (mean  SD; stress:
3.01  1.53 nmol/l; neutral: 3.13  1.55 nmol/l) and an
anticipatory stress response.
3.1.2. Negative affect
We first assessed whether baseline negative affect scores
differed for the two menstrual cycle phases. As expected,
baseline scores (15 min after arrival) were higher during the
luteal phase (mean  SD, 12.11  2.20) than during the folli-
cular phase (mean  SD, 10.96  1.11) (t27 = 2.78, P < 0.01).
Second, we tested whether stress induction reduced affect
ratings as intended. Negative affect ratings were higher after
the aversive movie clips than after the neutral control clips
(mean  SD, 14.09  3.42 and 12.73  2.68, respectively,
t27 = 2.66, P = 0.01) (Fig. 2). However, this increase in negative
affect did not differ significantly between menstrual cycle
phases (t < 1).
3.1.3. Heart rate
Overall, heart rate was higher during the aversive movie than
during the neutral control movie (t23 = 2.48, P < 0.05). As
predicted, heart rate increased slightly stronger during the
late luteal (mean  SD, from 62.97  10.48 BPM to
67.90  11.77 BPM) than during the late follicular phase
(from 61.69  8.77 BPM to 64.14  10.82 BPM) (t23 = 1.69,
P = 0.05, one-tailed (Fig. 2), and this is in line with previous
research (Tersman et al., 1991). Thus, the heart rate data
suggest that our stress induction method was successful, and
furthermore imply that stress responsiveness is somewhat
larger during the late luteal phase.
Menstrual cycle and stress 51

3.1.4. Reaction times
Reaction times to the onsets of fixation blocks (mean  SD,
675.5  260 ms) were not affected by stress induction and
did not differ between the two cycle phases. In addition, on-
line eye tracking was used to check whether the movies and
emotional expressions were attentively viewed. Both out-
comes confirmed that all women were fully alert during the
experimental procedure.
3.2. Imaging results
Brain regions generally involved in the processing of faces
with dynamically changing emotional expressions were iden-
tified first. This main effect of face presentation (i.e., emo-
tion blocks > fixation blocks) revealed widespread activation
in the occipital cortex (extending into the inferior temporal
cortex and including the fusiform gyrus), the amygdala, the
hippocampus, and the precentral gyrus extending into the
frontal eye fields (Table 1). Next, we tested whether any of
these regions exhibited a differential response to the three
facial expressions. No such main effect of facial expressions
was found in our main regions of interest (amygdala and
mPFC; both P > 0.05, SVC). Moreover, no interactions
between emotional expression and menstrual cycle phase
or stress (stress induction versus control) were found in these
brain regions (all P > 0.05, SVC). Therefore, remaining ana-
lyses were performed discarding the facial expression factor.
Subsequently, main effects of the factors menstrual cycle
phase and stress induction, and their interactions, were inves-
tigated. Neither the factor menstrual cycle phase nor stress
induction revealed main effects in the amygdala or mPFC (all
P > 0.05, SVC). However, an interaction between the factors
menstrual cycle phase and stress induction was found in the
amygdala (Figs. 1 and 2). Further testing showed a menstrual
cycle effect in the neutral condition, with larger amygdala
responses in the late luteal phase than late follicular phase.
Furthermore, the menstrual cycle phase by stress induction
interaction was carried by smaller amygdala responses after
stress induction during the late luteal phase (Fig. 2). Although
there was no main effect of face presentation in mPFC, this
region exhibited a similar menstrual cycle phase by stress
induction interaction (Figs. 1 and 2): the mPFC responses in
the neutral condition and after stress induction were reversed
across the two menstrual cycle phases.
To investigate whether this stress-induced decrease in
amygdala and mPFC responses could be mediated by allo-
pregnanolone, a correlation analysis was performed. The
increase in allopregnanolone levels from the follicular to
the luteal phase correlated with the stress-induced decrease
in amygdala and mPFC responses during the late luteal phase.
Specifically, the larger the increase in allopregnanolone
concentration across the menstrual cycle was, the smaller
the amygdala and mPFC responses were after stress induction
in the late luteal phase (Table 1).

Table 1 Local maxima for significant areas of activation for the main effects, menstrual cycle phase  stress induction interaction
and correlation analysis.

Side x y z T Value
Main effect of dynamic facial expressions > fixation baseline
Amygdala R 20 4 16 4.43 *
Amygdala L 20 4 18 3.20 *
Inferior occipital cortex R/L 44 76 10 29.98 **
Precentral gyrus R 52 4 48 9.42 **
Precentral gyrus L 52 0 50 7.33 **
Auditory association area R 50 40 10 7.89 **
Hippocampus R 22 26 6 7.41 **
Hippocampus L 20 28 6 7.61 **
Superior parietal cortex R 32 54 56 5.21 **
Superior temporal gyrus R 66 38 18 4.74 **
Menstrual cycle  stress induction interaction effect
Amygdala R 28 6 22 3.28 *
Medial prefrontal cortex R 14 50 6 3.88 *
Medial prefrontal cortex L 10 48 4 4.11 *
Luteal phase stress induction effect (neutral > stress)
Amygdala R 26 6 24 3.15 *
Menstrual cycle effect (late luteal > late follicular neutral condition)
Amygdala R 28 6 22 3.36 *
Correlational analysis (negative correlation: luteal phase stress induction effect with D allopregnanolone)
Amygdala R 22 0 10 3.91 *
Amygdala L 22 2 10 3.62 *
Medial prefrontal cortex R 4 56 12 4.69 *
Abbreviations: R: right; L: left.
*
P < 0.05, FWE rate corrected for the reduced search volume (i.e., amygdala or mPFC).
**
P < 0.05, FWE rate corrected for the entire brain.
52
Figure 1 Menstrual cycle phase  stress induction interaction in amygdala and mPFC. The figure shows coronal slices at y =
the amygdala and at y = 48 for the mPFC (P < 0.001, uncorrected).
4. Discussion
The present study investigated menstrual cycle-dependent
changes in neural mechanisms related to stress-sensitivity by
comparing the effects of acute stress on brain areas under-
lying emotion regulation in the late luteal with those in the
late follicular phase. Our results confirm increased stress-
sensitivity in the late luteal phase, as measured by the
highest negative affect scores and a stress-induced heart
rate increase. Certainly, heart rates after stress induction
were still quite low. However, previous studies using the same
stress induction procedure observed similar heart rates and
showed clear stress-induced differences in brain activity
(Henckens et al., 2009; Qin et al., 2009; van Marle et al.,
2009). Thus, our results suggest a successful stress induction,
although the overall heart rate is lower when compared to
other stress studies using for example the Trier Social Stress
Test, in which the participants are actively involved in
executing a task. Since no significant differences in cortisol
levels were observed, we suggest that the effects of stress
found here were mainly carried by immediate sympathetic
and (nor) adrenergic activation. Neuroimaging data revealed
an unexpected non-linear relationship between evoked
stress responses and subsequent neural activity in the amyg-
dala and the mPFC in response to biologically salient stimuli.
Data from the neutral conditions replicate earlier findings of
stronger responses in these regions in the late luteal phase
(Protopopescu et al., 2005). In contrast, this menstrual cycle-
dependent effect was inverted after stress induction in this
phase of the menstrual cycle. Importantly, this inverted
phasic amygdala and mPFC response was positively corre-
lated with changes in allopregnanolone concentration across
the menstrual cycle. These surprising findings implicate that
psychological stress mediates neural activity in a set of brain
regions implicated in emotion regulation in a cycle-depen-
dent manner, potentially via a mechanism that involves the
neuroactive steroid allopregnanolone.
A number of previous studies point towards a positive
association between emotional arousal and neural activity in
the amygdala and other regulatory circuits. For instance,
such activity has been shown to increase after negative mood
induction (Wang et al., 2006), viewing of both positive and
negative emotional expressions (Yang et al., 2002), and to
L. Ossewaarde et al.
6 for
correlate positively with evoked heart rate (Davis and Wha-
len, 2001; Critchley et al., 2005). Yet, negative relationships
have also been reported (Wang et al., 2005; Canli and Lesch,
2007). Decreases in phasic activity, as measured using BOLD—
fMRI, have been argued to occur against a background of
slowly modulated states of prolonged activation, which are
referred to as tonic activation states (Canli and Lesch, 2007).
The reason why such effects occur is that the BOLD signal
during control conditions, which serves as the baseline for
estimating the phasic response magnitude in BOLD—fMRI,
may itself be elevated in such states of tonic activation,
thus diminishing the difference between control and experi-
mental conditions (Wang et al., 2005; Canli and Lesch, 2007).
Hence, the stress-induced decrease in phasic response mag-
nitudes in the amygdala and medial prefrontal cortex during
the late luteal phase may actually be a consequence of
tonically increased activity, which has been observed pre-
viously in the amygdala of subjects with increased epigenetic
vulnerability for depression (Canli et al., 2006). The sug-
gested increase in tonic activity might be provoked by both
stress and menstrual cycle phase. However, with standard
fMRI methods as used in this study it is not possible to assess
changes in tonic activity. Merely the phasic BOLD—fMRI
responses during the dynamic facial expression task could
be measured. Therefore, an increase in tonic activity during
the late luteal phase under stress condition cannot directly
be confirmed in this study. This hypothesis can only be tested
by performing another menstrual cycle study using quanti-
tative CBF measurements such as positron emission tomo-
graphy (PET) or arterial spin labeling (ASL). Thus, future
research will be needed to resolve this issue. Regardless,
it is important to note that a tonic increase in brain activity is
central to the most prominent and widely used neurophysio-
logic model of the central stress response (Aston-Jones and
Cohen, 2005) and smaller phasic responses, as observed here,
is a central feature of this model.
The magnitude of this stress-induced reduction in phasic
neural responses in the luteal phase correlated with men-
strual cycle-related changes in salivary allopregnanolone,
with the lowest phasic response magnitudes in those women
with the strongest luteal phase increase of allopregnanolone.
Allopregnanolone modulates neural activity by potentiating
neuronal inhibition through action on the GABA-A receptor
Menstrual cycle and stress
Figure 2 Behavioral, physiological, and fMRI results during the
follicular and luteal phase for the stress and neutral condition.
(A) Relative differences on PANAS negative affect scores (for
visualization purposes, mean individual scores across all test
sessions were subtracted from every individual score at each
single test session): main effect stress induction: stress > neu-
neutral. Mean scores  SEM for the different test sessions: folli-
cular neutral: 12.18  0.43; follicular stress: 13.68  3.36;
luteal neutral: 13.29  0.77; luteal stress: 14.50  0.92. (B)
Relative differences in heart rate (for visualization purposes,
same calculation as A): main effect stress induction:
stress > neutral; menstrual cycle  stress induction interaction;
Stress induction effect in luteal phase: stress > neutral; men-
53
(Majewska et al., 1986). Furthermore, recent animal studies
have demonstrated that allopregnanolone modulates tonic
inhibition by action on d subunit-containing GABA-A recep-
tors, which influences anxiety and seizure susceptibility
(Maguire et al., 2005). In agreement, increases in saccadic
eye velocity during the luteal phase have been shown in
healthy women. Saccadic eye velocity is associated with
reduced GABAergic inhibition on saccade-related cells in
the superior colliculus, and has also been shown to be related
to the severity of symptoms in PMDD (Sundström and Bäck-
ström, 1998). Therefore, the hypothesized stress-induced
increase in tonic activation during the late luteal phase
may result from the influence of allopregnanolone on tonic
inhibition by action on the GABA-system.
Our findings of a correlation between allopregnanolone
and neural responses in emotion regulation areas converge
with a recent study on physically healthy postmenopausal
women receiving sequential hormone therapy with estrogen
and progesterone (Andréen et al., 2006). This study showed
increases in allopregnanolone during the progesterone treat-
ment phase. Subsequently, these women developed negative
mood symptoms, which were related to the allopregnanolone
concentration in an inverted U-shaped fashion. Luteal phase
range allopregnanolone concentrations coincided with
increased negative mood symptoms, but supraphysiological
concentrations were related to lower negative mood symp-
toms (Andréen et al., 2006). Earlier studies have shown that
allopregnanolone increases substantially during stress (Serra
et al., 2000; Droogleever Fortuyn et al., 2004). Therefore, an
alternative explanation of our neuroimaging findings could be
that in the situation in which concentrations of neuroactive
steroids are likely highest (i.e., after stress induction in the
luteal phase), a similar inverted U-shaped relation between
allopregnanolone and BOLD—fMRI responses in the amygdala
and mPFC may give rise to a decreased neural response.
However, we think that this explanation is less plausible,
because the present study found no reduction in negative
affect in the late luteal phase under stress, and even showed
that physiological stress responses in the luteal phase are
stronger than those in the follicular phase. Therefore, we
suggest that our results are explained best by a stress-
induced increase in tonic activity during the luteal phase.
The influence of gonadal hormones on brain areas involved
in emotion regulation has been shown in a number of previous
studies. These revealed that menstrual cycle phase, as well
as progesterone, affects activity in the amygdala and mPFC
(Goldstein et al., 2005; Dreher et al., 2007; van Wingen
et al., 2007, 2008; Derntl et al., 2008). Moreover, women
strual cycle effect in stress condition: luteal > follicular. Mean
scores  SEM for the different test sessions: follicular neutral:
61.69  1.69; follicular stress: 64.14  2.08; luteal neutral:
62.97  2.02; luteal stress: 67.90  2.27. (C) Menstrual cycle-
by-stress interaction in the amygdala based on activity at the
peak voxel: menstrual cycle effect in neutral condition:
luteal > follicular. Stress induction effect in luteal phase: neu-
tral > stress. (D) Menstrual cycle-by-stress interaction in the
mPFC based on activity at the peak voxel. No significant simple
effects. Error bars indicate SEM. Important to note: due to global
scaling of the imaging data the arbitrary units do not indicate
negative activations.
54
with PMDD appear more sensitive to those hormonal fluctua-
tions, and show enhanced amygdala and reduced mPFC
activity during the late luteal phase (Protopopescu et al.,
2008), a pattern similar to patients with mood and anxiety
disorders (Rauch et al., 2003; Drevets et al., 2008). The
present study extends these findings by showing that, in
healthy women, psychological stress exaggerates such men-
strual cycle effects in brain regions involved in emotion
regulation. This finding suggests that psychological stress
may promote the development of premenstrual symptoms.
Furthermore, it may have broader relevance for understand-
ing the role of psychological stress in the etiology of psy-
chiatric disorders related to fluctuations in the hormonal
milieu across the reproductive cycle. Future research into
these issues is warranted, and should take these factors into
account.
In conclusion, the present study reveals a neural mechan-
ism that may mediate increased stress-sensitivity during the
late luteal phase. Surprisingly, increased stress-sensitivity, as
indicated by a mild increase in heart rate responses to a
stressor, was accompanied by lower phasic activity in the
amygdala and mPFC, which might imply a stress-induced
state of enhanced tonic activity in these areas. Moreover,
our results suggest that this effect appears to be mediated by
changes in neural excitability associated with fluctuations of
allopregnanolone across the menstrual cycle. These findings
may help gain more insight in the pathogenesis of PMDD and
other mood disturbances associated with hormonal fluctua-
tions across the reproductive life cycle.
Conflict of interest
None declared.
Acknowledgements
This work was supported by grant numbers 918.66.613 and
451.07.019 from the Netherlands Organization for Scientific
Research (NWO) and grant no. 14x-11198 from the Swedish
research council.
References
Andréen, L., Sundström-Poromaa, I., Bixo, M., Nyberg, S., Bäck-
ström, T., 2006. Allopregnanolone concentration and mood–—a
bimodal association in postmenopausal women treated with oral
progesterone. Psychopharmacology (Berl). 187, 209—221.
Aston-Jones, G., Cohen, J.D., 2005. An integrative theory of locus
coeruleus-norepinephrine function: adaptive gain and optimal
performance. Annu. Rev. Neurosci. 28, 403—450.
Bäckström, T., Sanders, D., Leask, R., Davidson, D., Warner, P.,
Bancroft, J., 1983. Mood, sexuality, hormones, and the menstrual
cycle. II. Hormone levels and their relationship to the premenstr-
ual syndrome. Psychosom. Med. 45, 503—507.
Belelli, D., Lambert, J., 2005. Neurosteroids: endogenous regulators
of the GABAA receptor. Nat. Rev. Neurosci. 6, 565—575.
Canli, T., Lesch, K.P., 2007. Long story short: the serotonin trans-
porter in emotion regulation and social cognition. Nat. Neurosci.
10, 1103—1109.
Canli, T., Qiu, M., Omura, K., Congdon, E., Haas, B.W., Amin, Z.,
Herrmann, M.J., Constable, R.T., Lesch, K.P., 2006. Neural corre-
lates of epigenesis. Proc. Natl. Acad. Sci. U.S.A. 103, 16033—16038.
L. Ossewaarde et al.
Critchley, H.D., Rotshtein, P., Nagai, Y., O’Doherty, J., Mathias, C.J.,
Dolan, R.J., 2005. Activity in the human brain predicting differ-
ential heart rate responses to emotional facial expressions.
Neuroimage 24, 751—762.
Davis, M., Whalen, P.J., 2001. The amygdala: vigilance and emotion.
Mol. Psychiatry 6, 13—34.
de Zwart, J.A., van Gelderen, P., Golay, X., Ikonomidou, V.N., Duyn,
J.H., 2006. Accelerated parallel imaging for functional imaging of
the human brain. NMR Biomed. 19, 342—351.
Derntl, B., Windischberger, C., Robinson, S., Lamplmayr, E., Kryspin-
Exner, I., Gur, R.C., Moser, E., Habel, U., 2008. Facial emotion
recognition and amygdala activation are associated with men-
strual cycle phase. Psychoneuroendocrinology 33, 1031—1040.
Dreher, J.C., Schmidt, P.J., Kohn, P., Furman, D., Rubinow, D.,
Berman, K.F., 2007. Menstrual cycle phase modulates reward-
related neural function in women. Proc. Natl. Acad. Sci. U.S.A.
104, 2465—2470.
Drevets, W.C., Price, J.L., Furey, M.L., 2008. Brain structural and
functional abnormalities in mood disorders: implications for neu-
rocircuitry models of depression. Brain Struct. Funct. 213, 93—118.
Droogleever Fortuyn, H.A., van Broekhoven, F., Span, P.N., Back-
strom, T., Zitman, F.G., Verkes, R.J., 2004. Effects of PhD exam-
ination stress on allopregnanolone and cortisol plasma levels and
peripheral benzodiazepine receptor density. Psychoneuroendo-
crinology 29, 1341—1344.
Ekman, P., Friesen, W.V., 1976. Pictures of Facial Affect. Consulting
Psychologists Press, Palo Alto, CA.
Epperson, C.N., Pittman, B., Czarkowski, K.A., Stiklus, S., Krystal,
J.H., Grillon, C., 2007. Luteal-phase accentuation of acoustic
startle response in women with premenstrual dysphoric disorder.
Neuropsychopharmacology 32, 2190—2198.
Fernández, G., Weis, S., Stoffel-Wagner, B., Tendolkar, I., Reuber, M.,
Beyenburg, S., Klaver, P., Fell, J., de Greiff, A., Ruhlmann, J.,
Reul, J., Elger, C.E., 2003. Menstrual cycle-dependent neural
plasticity in the adult human brain is hormone, task, and region
specific. J. Neurosci. 23, 3790—3795.
Gold, P.W., Drevets, W.C., Charney, D.S., 2002. New insights into the
role of cortisol and the glucocorticoid receptor in severe depres-
sion. Biol. Psychiatry 52, 381—385.
Goldstein, J.M., Jerram, M., Poldrack, R., Ahern, T., Kennedy, D.N.,
Seidman, L.J., Makris, N., 2005. Hormonal cycle modulates arou-
sal circuitry in women using functional magnetic resonance ima-
ging. J. Neurosci. 25, 9309—9316.
Griswold, M.A., Jakob, P., Heidemann, R., Nittka, M., Jellus, V.,
Wang, J., Kiefer, B., Haase, A., 2002. Generalized autocalibrating
partially parallel acquisitions (GRAPPA). Magn. Reson. Med. 47,
1202—1210.
Henckens, M.J., Hermans, E.J., Pu, Z., Joëls, M., Fernández, G.,
2009. Stressed memories: how acute stress affects memory for-
mation in humans. J. Neurosci. 29, 10111—10119.
Herman, J.P., Ostrander, M.M., Mueller, N.K., Figueiredo, H., 2005.
Limbic system mechanisms of stress regulation: hypothalamo—
pituitary—adrenocortical axis. Prog. Neuropsychopharmacol.
Biol. Psychiatry 29, 1201—1213.
Kask, K., Gulinello, M., Bäckström, T., Geyer, M.A., Sundström-
Poromaa, I., 2008. Patients with premenstrual dysphoric disorder
have increased startle response across both cycle phases and
lower levels of prepulse inhibition during the late luteal phase of
the menstrual cycle. Neuropsychopharmacology 33, 2283—2290.
Kumari, V., Corr, P.J., 1998. Trait anxiety, stress and the menstrual
cycle: effects on Raven’s Standard Progressive Matrices test.
Pers. Indiv. Differ. 24, 615—623.
Lundqvist, D., Flykt, A., Öhman, A., 1998. The Karolinska Directed
Emotional Faces. Karolinska Institute, Stockholm.
Maguire, J.L., Stell, B.M., Rafizadeh, M., Mody, I., 2005. Ovarian
cycle-linked changes in GABA(A) receptors mediating tonic inhi-
bition alter seizure susceptibility and anxiety. Nat. Neurosci. 8,
797—804.
Menstrual cycle and stress
Majewska, M.D., Harrison, N.L., Schwartz, R.D., Barker, J.L., Paul,
S.M., 1986. Steroid hormone metabolites are barbiturate-
like modulators of the GABA receptor. Science 232 (4753),
1004—1007.
Marinari, K.T., Leshner, A.I., Doyle, M.P., 1976. Menstrual cycle status
and adrenocortical reactivity to psychological stress. Psychoneur-
oendocrinology 1, 213—218.
Moldovanova, I., Schroeder, C., Jacob, G., Hiemke, C., Diedrich, A.,
Luft, F.C., Jordan, J., 2008. Hormonal influences on cardiovas-
cular norepinephrine transporter responses in healthy women.
Hypertension 51, 1203—1209.
Moos, R.H., 1968. The development of a menstrual distress ques-
tionnaire. Psychosom. Med. 30, 853—867.
Phelps, E.A., LeDoux, J.E., 2005. Contributions of the amygdala to
emotion processing: from animal models to human behavior.
Neuron 48, 175—187.
Protopopescu, X., Pan, H., Altemus, M., Tuescher, O., Polanecsky, M.,
McEwen, B., Silbersweig, D., Stern, E., 2005. Orbitofrontal
cortex activity related to emotional processing changes across
the menstrual cycle. Proc. Natl. Acad. Sci. U.S.A. 102, 16060—
16065.
Protopopescu, X., Tuescher, O., Pan, H., Epstein, J., Root, J., Chang,
L., Altemus, M., Polanecsky, M., McEwen, B., Stern, E., Silbers-
weig, D., 2008. Toward a functional neuroanatomy of premenstr-
ual dysphoric disorder. J. Affect. Disord. 108, 87—94.
Purdy, R.H., Moore Jr., P.H., Rao, P.N., Hagino, N., Yamaguchi, T.,
Schmidt, P., Rubinow, D.R., Morrow, A.L., Paul, S.M., 1990.
Radioimmunoassay of 3 alpha-hydroxy-5 alpha-pregnan-20-one
in rat and human plasma. Steroids 55, 290—296.
Qin, S., Hermans, E.J., van Marle, H.J., Luo, J., Fernández, G., 2009.
Acute psychological stress reduces working memory-related
activity in the dorsolateral prefrontal cortex. Biol. Psychiatry
66, 25—32.
Rauch, S.L., Shin, L.M., Wright, C.I., 2003. Neuroimaging studies of
amygdala function in anxiety disorders. Ann. N. Y. Acad. Sci. 985,
389—410.
Sato, W., Kochiyama, T., Yoshikawa, S., Naito, E., Matsumura, M.,
2004. Enhanced neural activity in response to dynamic facial
expressions of emotion: an fMRI study. Brain Res. Cogn. Brain
Res. 20, 81—91.
Serra, M., Pisu, M.G., Littera, M., Papi, G., Sanna, E., Tuveri, F.,
Usala, L., Purdy, R.H., Biggio, G., 2000. Social isolation-induced
decreases in both the abundance of neuroactive steroids
and GABA(A) receptor function in rat brain. J. Neurochem. 75,
732—740.
Shansky, R.M., Glavis-Bloom, C., Lerman, D., McRae, P., Benson, C.,
Miller, K., Cosand, L., Horvath, T.L., Arnsten, A.F., 2004. Estrogen
mediates sex differences in stress-induced prefrontal cortex
dysfunction. Mol. Psychiatry 9, 531—538.
Sheehan, D.V., Lecrubier, Y., Sheehan, K.H., Amorim, P., Janavs, J.,
Weiller, E., Hergueta, T., Baker, R., Dunbar, G.C., 1998.
The Mini-International Neuropsychiatric Interview (M.I.N.I.):
the development and validation of a structured diagnostic psy-
chiatric interview for DSM-IV and ICD-10. J. Clin. Psychiatry 59
(Suppl. 20), 22—33.
Steiner, M., Dunn, E., Born, L., 2003. Hormones and mood: from
menarche to menopause and beyond. J. Affect. Disord. 74,
67—83.
55
Sundström, I., Bäckström, T., 1998. Patients with premenstrual
syndrome have decreased saccadic eye velocity compared to
control subjects. Biol. Psychiatry 44, 755—764.
Sveindottir, H., Bäckström, T., 2000. Prevalence of menstrual cycle
symptom cyclicity and premenstrual dysphoric disorder in a ran-
dom sample of women using and not using oral contraceptives.
Acta Obstet. Gynecol. Scand. 79, 405—413.
Talairach, J., Tournoux, P., 1988. Co-planar Stereotaxic Atlas of the
Human Brain: 3-Dimensional Proportional System–—An Approach
to Cerebral Imaging. Thieme, New York.
Tersman, Z., Collins, A., Eneroth, P., 1991. Cardiovascular responses
to psychological and physiological stressors during the menstrual
cycle. Psychosom. Med. 53, 185—197.
Turkmen, S., Lundgren, P., Birzniece, V., Zingmark, E., Bäckström, T.,
Johansson, I.M., 2004. 3beta-20beta-dihydroxy-5alpha-pregnane
(UC1011) antagonism of the GABA potentiation and the learning
impairment induced in rats by allopregnanolone. Eur. J. Neurosci.
20, 1604—1612.
van Marle, H.J., Hermans, E.J., Qin, S., Fernández, G. 2009. From
specificity to sensitivity: how acute stress affects amygdala
processing of biologically salient stimuli. Biol. Psychiatry, [Epub
ahead of print], doi:10.1016/j.biopsych.2009.05.014.
van Wingen, G., van Broekhoven, F., Verkes, R.J., Petersson, K.M.,
Bäckström, T., Buitelaar, J., Fernández, G., 2007. How proges-
terone impairs memory for biologically salient stimuli in healthy
young women. J. Neurosci. 27, 11416—11423.
van Wingen, G.A., van Broekhoven, F., Verkes, R.J., Petersson, K.M.,
Bäckström, T., Buitelaar, J.K., Fernández, G., 2008. Progesterone
selectively increases amygdala reactivity in women. Mol. Psy-
chiatry 13, 325—333.
Wang, J., Rao, H., Wetmore, G.S., Furlan, P.M., Korczykowski, M.,
Dinges, D.F., Detre, J.A., 2005. Perfusion functional MRI reveals
cerebral blood flow pattern under psychological stress. Proc.
Natl. Acad. Sci. U.S.A. 102, 17804—17809.
Wang, L., LaBar, K.S., McCarthy, G., 2006. Mood alters amygdala
activation to sad distractors during an attentional task. Biol.
Psychiatry 60, 1139—1146.
Wang, M., Seippel, L., Purdy, R.H., Bäckström, T., 1996. Relationship
between symptom severity and steroid variation in women with
premenstrual syndrome: study on serum pregnenolone, pregne-
nolone sulfate, 5 alpha-pregnane-3,20-dione and 3 alpha-
hydroxy-5 alpha-pregnan-20-one. J. Clin. Endocrinol. Metab.
81, 1076—1082.
Watson, D., Clark, L.A., Tellegen, A., 1988. Development and valida-
tion of brief measures of positive and negative affect: the PANAS
scales. J. Pers. Soc. Psychol. 54, 1063—1070.
Weis, S., Hausmann, M., Stoffers, B., Vohn, R., Kellermann, T.,
Sturm, W., 2008. Estradiol modulates functional brain organiza-
tion during the menstrual cycle: an analysis of interhemispheric
inhibition. J. Neurosci. 28, 13401—13410.
Worsley, K., Marrett, S., Neelin, P., Vandal, A.C., Friston, K.J., Evans,
A.C., 1996. A unified statistical approach for determining sig-
nificant signals in images of cerebral activation. Hum. Brain
Mapp. 4, 58—73.
Yang, T.T., Menon, V., Eliez, S., Blasey, C., White, C.D., Reid, A.J.,
Gotlib, I.H., Reiss, A.L., 2002. Amygdalar activation associated
with positive and negative facial expressions. Neuroreport 13,
1737—1741.