Hormones and Behavior 47 (2005) 267 – 271
www.elsevier.com/locate/yhbeh
Fear recognition across the menstrual cycle
Rebecca Pearson, Michael B. Lewis*
School of Psychology, Cardiff University, Park Place, Cardiff, CF10 3AT, UK
Received 28 September 2004; revised 19 November 2004; accepted 23 November 2004
Available online 16 January 2005
Abstract
This study assesses the mediating role of stage of menstrual cycle in the recognition of emotional expressions. It was hypothesised that
fear recognition ability would be stronger at high-oestrogen stages of the menstrual cycle. The accuracy of recognising emotional expressions
was compared across 50 women who were at different stages of their menstrual cycle. It was found that accuracy to recognise emotions was
significantly affected by the interaction between stages of the menstrual cycle and the emotion being displayed. Further analysis revealed that
for the emotion expression of fear alone, participants were significantly more accurate at the preovulatory surge (highest oestrogen levels)
than at menstruation (oestrogen levels at lowest point). The results have implications for the processes that underlie fear processing and a
possible insight into the sexual dimorphism of this ability and conditions that show variations in fear recognition (e.g., autism, Turner
syndrome).
D 2005 Elsevier Inc. All rights reserved.
Keywords: Fear; Oestrogen; Menstrual cycle; Sexual dimorphism; Autism; Depression
Introduction
Cognitions change across the menstrual cycle: for
example, Penton-Voak and Perrett (2000) found that during
the preovulatory and ovulatory phases of the cycle, women
prefer more masculinised faces than at other stages. At these
high oestrogen phases of the menstrual cycle, women are at
their most fertile and, it can be argued, it makes evolutionary
sense to have an enhanced sensitivity to reproductively
relevant stimuli. Macrae et al. (2002) suggest that when
conception risk is high, women are primed by hormones to
make more efficient social cognitive judgements. They
found that women made more categorical and stereotypical
judgements about men at ovulation relative to menstruation.
This implies that women are more efficient at, and may have
a heightened ability for, person perception and social
cognition at times when they are most fertile.
Social competence is itself a sexually dimorphic skill
with women showing a superior empathizing ability
according to Baron-Cohen’s (2003) definition of the female
* Corresponding author.
E-mail address: [email protected] (M.B. Lewis).
0018-506X/$ - see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.yhbeh.2004.11.003
brain. Baron-Cohen states that women show earlier development and superior theory of mind abilities (i.e., the
developed understanding that other humans have beliefs
and a point of view different to one’s own). This ability may
be linked to sexually dimorphic processing of facial
expressions in which women generally show a superior
ability (Hall et al., 1999). Women are both better encoders
of basic emotions (Kilgore, 2000) and more complex
emotions and mental states (Baron-Cohen, 2003). It is
argued below that fear recognition is at the heart of this
sexual dimorphism.
Responses to a threatening environment are often important to the survival of species and the amygdala is
instantly activated when threat, such as a fearful face, is
detected (Morris et al., 1998). Despite showing activation to
all emotional stimuli, the amygdala shows preferential
activation to fearful faces (Morris et al., 1996) and brain
damage to this region results in the selective impairment of
fear recognition (Adolphs et al., 1995). Calder et al. (2001)
provide a review of the relationship between fear and the
amygdale.
Skuse (2003) argues that the neural circuits based around
the amygdala have adapted to enable higher social cognitive
268
R. Pearson, M.B. Lewis / Hormones and Behavior 47 (2005) 267–271
functions such as a theory of mind. Skuse proposes that the
recognition of the facial expression of fear correlates with
theory of mind ability. He further suggests that this is
because, unlike the direct threat of anger, for example,
understanding why a fearful face signals threat requires us to
interpret the world through someone else’s perspective. To
discover what the threat is, we need to look outside the
person and see the subject of their fear. Successful
interpretation of the meaning of a fearful face, thus, requires
second-order representations and a theory of mind. It has
been argued that these cognitive abilities separate humans
from other species (Tomasello, 1999).
Fear recognition, unlike any of the other emotions,
correlates with face recognition memory in women but not
men (Campbell et al., 2002). This suggests that the specific
neural mechanisms involved in fear recognition may be
sexually dimorphic. Neuroimaging studies enable researchers to identify the pattern of structural and functional sexual
dimorphism of the neural network. The main areas of sexual
differentiation involve hemispheric lateralisation: ERP data
show men to have greater right hemisphere activity when
processing emotions (Everhart et al., 2001). Morris et al.
(1998) demonstrated that the right amygdala is associated
with automatic and even unconscious processing of fear
whereas the left amygdala is involved in a full analysis of
the face and has links to higher regions. A right amygdala
advantage in men may therefore reflect the tendency to rely
on an initial automatic response, perhaps reminiscent of the
ancient fight-or-flight reaction.
Women, however, exhibit a differential pattern of
activation in the left amygdala and subsequent regions such
as the prefrontal cortex and the hippocampus when
presented with fearful faces. Women, relative to men,
exhibit increased habituation in the left amygdala followed
by increased activity in the hippocampus (Campbell et al.,
2002; Kilgore, 2000). The neural differences may reflect an
adaptive female differentiation of the ancient neural
bwarning systemQ. Increased habituation may reflect an
inhibitory mechanism controlling the initial fight-or-flight
response and evaluating the social context. Successfully
attributing salience and value to the emotional stimuli
requires careful balance of the amygdala and higher cortical
regions (Skuse, 2003). This involves accessing emotional
memory (the hippocampus) and the evaluative capabilities
of the left prefrontal cortex (Ledoux, 1995). As discussed
above, these regions show more activation in women’s
response to fear and women only show an association of
fear recognition with memory for faces placing a special
role for the hippocampus in women.
Women compared to men take more notice of, and assign
more meaning to, emotions in others (Baron-Cohen, 2003).
This may reflect the specific evolutionary pressures on
women. For example a fight-or-flight response would not be
an appropriate reaction to a young child’s fearful expression;
a mother needs to evaluate meanings so that she can
distinguish realistic and unrealistic fears in the child, as she
is responsible for teaching them about the world. As
hunters, men would primarily meet fearful expressions in
situations where a rapid fight-or-flight reaction provides the
most promising survival tactic.
It is hypothesised that oestrogen may play a mediating
role in this female adaptation and may explain the structural
and functional differences discussed above. Oestrogen
receptors have been identified in the amygdala (Osterlund
and Hurd, 2001) and more prominently in the hippocampus
and the corpus callosum (Fitch and Denenberg, 1998).
Oestrogens acting on these sites can change the neurochemical and physical structures of cells (McEwen et al.,
1997) and cycling estrus hormones in the female rat are
known to remodel hippocampal cells (Woolley et al., 1990).
It appears then that oestrogen facilitates increased functioning of these areas. As discussed before, the transient effects
of hormones on these brain regions may explain the female
brain acting more bfemaleQ at high oestrogen stages of the
menstrual cycle.
Further evidence of the link between hormones and fear
recognition comes from clinical disorders, such as autism,
that appear to have very specific impairments in fear
recognition (Howard et al., 2000). Autism is a predominantly male disorder and has been conceptualised by
Baron-Cohen (2003) as the extreme male brain. Part of his
biological theory behind masculinisation or feminisation of
the brain is the level of male and female hormones the brain
is exposed to. Autism can be characterised as a severe lack
of active female hormones in the central nervous system.
This theory is further supported by evidence from studies
into Turner’s syndrome. Patients with this syndrome suffer
from ovarian dysgenesis and the subsequent failure of
endogenous oestrogen production (Collear et al., 2002).
Individuals with Turner’s syndrome are at an increased risk
of having autism (at least 200 times, Creswell and Skuse,
1999) and show impaired social abilities (Lawrence et al.,
2003). Like individuals with autism, individuals with
Turner’s syndrome have a specific deficit in fear recognition
(Lawrence et al., 2003). It would appear that without
oestrogen, the brains of individuals with Turner’s syndrome
fail to dfeminiseT resulting in some extreme-male or autisticlike characteristics. Further evidence for oestrogen’s key
role in fear recognition is that emerging evidence suggests
that individuals who have received early oestrogen replacement are less likely to show this fear recognition deficit
which appears to be so key in social cognition (Lawrence
et al., 2003).
While we have focused on the possible effect of oestrogen
here, it must be considered that an effect of menstrual stage
could also be accounted for in terms of other female
hormones such as progesterone. Such an effect would be
supported by work conducted by Smith et al. (1998), who
demonstrated how progesterone’s metabolite allopregnanolone is active in the hippocampus area.
The current experiment explored how accuracy of
recognition of six basic facial expressions of emotions
R. Pearson, M.B. Lewis / Hormones and Behavior 47 (2005) 267–271
varies across four stages of the menstrual cycle. It was
hypothesised that, specifically, fear-recognition ability will
be at its most accurate during high oestrogen phases,
particularly the preovulatory phase where oestrogen is at
its highest, and the ability will be at its lowest at low
oestrogen phases, particularly at menstruation when oestrogen is at its lowest.
Method
Participants
Fifty-three female undergraduate students from Cardiff
University volunteered to take part in the study in return for
course credit. The ages ranged from 18 to 22 and the mean
age was 20. All the participants that were included had a
regular menstrual cycle and were not taking any form of the
contraceptive pill. Three participants were excluded from
the study: two because of ambiguous menstrual cycles and
one due to computer failure. The participants were assigned
to one of four groups depending on the stage of their
menstrual cycle they were experiencing on the date of
testing. In order to track the stages of the menstrual cycle,
two key events from each participant were noted. If the
female reported she was currently menstruating or had been
in the last 2 days, she was automatically assigned to the
menstruation group (N = 12). The three remaining groups
were calculated by obtaining information via email from the
individuals of their subsequent onset of menses in return for
further course credit. This enabled calculation of ovulation,
which is taken to have occurred 14 days prior to
menstruation. Participants who were tested the day before
and 3 days after ovulation were assigned to the ovulating
group (i.e., 15 days to 11 days before the onset of
subsequent menses; N = 11). Participants who were tested
before ovulation were assigned to the preovulatory group
(N = 13) and participants who were tested after ovulation
were assigned to the luteal group (N = 14).
269
and that they had to decide how they thought the person
was feeling from the six options, which were on the
keyboard. Keys on a keyboard were labelled with stickers
with capital print of the six emotions. The position of each
emotion sticker on the keyboard was counterbalanced
across participants. They were told there would be six
practice faces followed by a break and then they would
enter the test phase. They were instructed to make their
response as soon as they had come to a decision. There was
an interstimulus interval of 1000 ms and each test item was
presented until the participant made a keyboard response.
After completion of the computerised test, the participants
were instructed to email to confirm the date of their next set
of menses.
Design
This was a two-way mixed design. The within-subject
independent variable was emotion and has six levels: fear,
happy, disgust, sad, surprise and anger. The betweensubjects independent variable is menstrual cycle stage and
has 4 levels: menses, preovulatory, ovulating and luteal.
There were two dependent variables. The first is accuracy
scores for correct recognition of each emotion out of a
maximum of 10. The second dependent variable was the
mean reaction time to make correct responses to each of the
emotional stimuli sets.
Results
The results for the accuracy scores for the six emotions
are shown in Fig. 1. This shows that variations in accuracy
for emotions over cycle stages are greatest for fear
recognition. Accuracy for fear recognition is greatest during
the pre-ovulation stage and at its lowest during menses.
Stimuli
The stimuli consisted of 6 practice and 60 test faces.
These were taken from the Ekman and Friesen (1976) set,
which has been used in emotion recognition studies as a
valid measure for over 20 years. The test consisted of 11
fearful, 11 sad, 11 happy, 11 angry, 11 disgusted and 11
surprised faces. One of each was presented in the practice
and the remaining 60 were presented on the screen in
randomised order.
Procedure
Participants were initially given a consent form and then
instructed to fill out the menstrual cycle questionnaire. They
were then told that several faces would appear on the screen
Fig. 1. Mean accuracy for recognition of the six different types of emotional
expressions by participants according to which stage of the menstrual
cycles that were at when tested. Accuracy of the fear faces was affected by
the menstrual cycle stage. Error bars show one standard error.
270
R. Pearson, M.B. Lewis / Hormones and Behavior 47 (2005) 267–271
A two-way mixed design ANOVA was performed on the
accuracy data. There was a significant main effect for
emotion [ F(5,230) = 36.764; P b 0.001]. The emotion-bystage interaction was also significant [ F(15,230) = 2.379;
P b 0.01]. Simple main effects analysis was performed with
planned pairwise comparisons. Effects of stage of cycle
were found for fear where performance at menses was
significantly ( P b 0.05) poorer than at pre-ovulation or
ovulating stages. Other comparisons were nonsignificant.
Although hormone levels were not assessed directly for
the participants, it is possible to have a general idea as to the
approximate level of oestrogen for each of the four groups
based on their stage of cycle reported. Using charts such as
that produced by Alliende (2002), a rough measure of
estrone glucronide for participants in each of the four
groups can be estimated. Participants would be likely on
average to have the following averages: the menses group
would have 29.8 nmol/l/day; the pre-ovulation group would
have 62.4 nmol/l/day; the ovulation group would have
53.7 nmol/l/day and luteal group would have 42.9 nmol/l/
day. These values for the four groups can be correlated with
the accuracy obtained for the fear faces. This correlation
reveals a strong positive relationship between performance
on the fearful faces and the level of estrone glucuronide (r =
0.987; P b 0.01).
A mixed-design ANOVA was conducted for mean
reactions. Although this revealed that happy faces could
be categorised faster than any other emotion faces ( P b
0.05), no other effects or interactions were significant.
Discussion
It can be concluded that there is a relationship between
fear recognition and stage of the menstrual cycle. This
finding supports the hypothesis regarding the level of
oestrogen accuracy and encoding of fearful faces. Accuracy
for fearful faces was highest during their highest oestrogen
phase (preovulatory surge) and was lowest at the lowest
oestrogen phase (menstruation). In fact, the pattern of
accuracy scores across all four stages matches the pattern
of oestrogen fluctuations across these stages as indicated by
the strong correlation. The fact that fear recognition is
differentially affected by stage of cycle suggests that there is
a distinct neural processing of fear. The pattern of results
also provides strong evidence for oestrogen’s (or some other
hormone’s) mediating role in this system.
It is, however, possible that the mechanism behind the
difference in performance can be explained in terms of
mood changes and associated cognitive differences during
the menstrual cycle (e.g., Davydov et al., 2004). If the
participants were suffering from premenstrual dysphoric
disorder or premenstrual syndrome, then it would be
expected that performance would be better during the
preovulatory and ovaluation periods than during the menses
period as has been shown in a variety of tasks (e.g., Mans
et al., 1999). This explanation, however, would not explain
why it is that the recognition of the fear emotion specifically
is affected by the menstrual cycle.
Previous studies have shown that, at menstruation, women’s cognitive abilities shift slightly towards a male dtypeT
brain (Sanders et al., 2002). The experiment above has
demonstrated that, at menstruation, participants show a
significant decrease in fear recognition. Impairments at this
skill at an extreme level characterise the proposed extreme
male brain and are said to shape autistic individuals (BaronCohen, 2003). The relationship demonstrated here between
this ability and hormones provides a new direction in research
into autism. Is extreme lack of oestrogen or progesterone
a potential cause? Could early oestrogen replacement act as
a protective factor as it appears to in Turner’s syndrome?
The association between hormones and fear recognition
also poses another question: if lack of oestrogen produces
the extreme male brain, it would follow that excess levels
will result in the extreme female brain. In his recent book,
Baron-Cohen (2003) leaves open the question: What is the
nature of the extreme female brain? Fear recognition may
provide the answer to this, just as individuals with an
extreme male brain deviate below the norm on this ability,
individuals with an extreme female brain may deviate above
the norm. Such individuals have been identified in a recent
study that shows that individuals with a history of a major
depressive episode show a selectively greater recognition of
fear (Bhagwager et al., 2004). The authors also demonstrate
that this abnormality was normalised through treatment with
selective serotonin reuptake inhibitors. Previously, emotion
recognition abilities have been found to correlate with
depression (Gur et al., 1992). Could oestrogen play a role in
the chemical imbalance of depressive disorder as has been
suggested by, for example, Angold et al. (1999)? Is the
predominantly female disorder of depression (2:1 female to
male ratio, Hammen, 1997) a characteristic of the extreme
female brain? This speculation requires further inquiry but
one potential outcome from this line of research would be a
test that could screen for depression (perhaps postnatal
depression) without having to ask explicitly about feelings.
Such a test could be administered as a test of hormonal
imbalance and require people to make judgements about
emotions expressed.
An implication of the current study for the normal female
population concerns the effects of methods of contraception
that alter levels of hormones might have. The present
findings would suggest that altering one’s hormonal levels
may subtly but significantly alter fear recognition ability.
When fear recognition has such strong associations with
social competence, interfering with such a process could
have negative implications. No individuals who were taking
the contraceptive pill were included in this study but an
important future direction would be to assess the ability of
those on different pills with the prediction that pills reducing
oestrogen levels would be associated with lower fear
recognition abilities.
R. Pearson, M.B. Lewis / Hormones and Behavior 47 (2005) 267–271
Conclusion
What can be concluded from the present findings is that
fear recognition, unlike recognition of other emotions, is
associated with hormone levels. Previous research has
placed fear recognition as a key social cognitive task. Oestrogen or progesterone then may find its role in the
feminisation of the brain, which may have been particularly
significant for the evolution of social cognition out of simple
instinctive responses to threat. Moreover, the transient
effects of these hormones across the menstrual cycle affirm
them as an active chemical in the adult female brain.
Seemingly, hormones’ active role in the brain not only shape
between-sex variations but within-sex variation. Extreme
variations may underlie cognitive and emotional disorders.
Further understanding of the complex relationship between
hormones, the brain and behaviour may aid understanding
into these disorders as well as the biologically caused
differences between the emerging dfeminineT and dmasculineT
cognitive sets.
References
Adolphs, R., Tranel, D., Damasio, H., Damasio, A.R., 1995. Fear and the
human amygdala. J. Neurosci. 15 (9), 5879 – 5891.
Alliende, M.E., 2002. Mean versus individual hormonal profiles in the
menstrual cycle. Fertil. Steril. 78, 90 – 95.
Angold, A., Costello, E.J., Erkanli, A., Worthman, C.M., 1999. Pubertal
changes in hormone levels and depression in girls. Psychol. Med. 29,
1043 – 1053.
Baron-Cohen, S., 2003. The Essential Difference, Men, Women and the
Extreme Male Brain. Penguin Press, London.
Bhagwager, Z., Caven, P.J., Goodwin, G.M., Harmer, C.J., 2004. Normalization of enhanced Fear recognition by acute SSRI treatment in
subjects with a previous history of depression. Am. J. Psychiatry 161,
166 – 168.
Calder, A.J., Lawrence, A.D., Young, A.W., 2001. Neuropsychology of fear
and loathing. Nat. Rev., Neurosci. 2, 352 – 363.
Campbell, R., Elgar, K., Kuntsi, J., Akers, R., Terstegge, J., Coleman, M.,
Skuse, D., 2002. The classification of dfearT from faces is associated
with face recognition skill in women. Neuropsychologia 40, 575 – 584.
Collear, M.L., Geffner, M.E., Kaufman, F.R., Buckingham, S., Hines, M.,
2002. Cognitive and behavioural characteristics of turner syndrome:
exploring a role of ovarian hormones in female sexual differentiation.
Horm. Behav. 41, 139 – 155.
Creswell, C.S., Skuse, D., 1999. Autism in association with Turner’s
syndrome: genetic implications for male vulnerability to pervasive
developmental disorders. Neurocase 5, 101 – 108.
Davydov, D.M., Shapiro, D., Goldstein, I.B., 2004. Moods in everyday
situations—effects of menstrual cycle, work and personality. J.
Psychosom. Res. 56, 27 – 33.
271
Ekman, P., Friesen, W.V., 1976. Pictures of Facial Affect. Consulting
Psychologists Press, Palo Alto.
Fitch, R.H., Denenberg, V.H., 1998. A role for ovarian hormones in sexual
differentiation of the brain. Behav. Brain Sci. 21 (3), 311 – 352.
Gur, R.C., Erwin, R.J., Gur, R.E., Zwil, A.S., Heimberg, C., Kraemer, H.C.,
1992. Facial emotion discrimination: II. Behavioural findings in
depression. Psychiatry Res. 42, 241 – 251.
Hall, C.W., Gaul, L., Kent, M., 1999. College students perception of facial
expressions. Percept. Mot. Skills 89, 763 – 770.
Hammen, C., 1997. Depression. Psychology Press, East Sussex.
Howard, M.A., Cowell, P.E., Boucher, J., Broks, P., Mayes, A., Farrant, A.,
Roberts, N., 2000. Convergent neuroanatomical and behavioural
evidence of an amygdale hypothesis of autism. NeuroReport 11,
2931 – 2935.
Kilgore, W.D.S., 2000. Sex differences in identifying the facial affect
of normal and mirror-reversed faces. J. Percept. Mot. Skills 91,
525 – 530.
Lawrence, K., Kuntsi, J., Coleman, M., Campbell, R., Skuse, D., 2003.
Face and emotion recognition deficits in Turner syndrome: a possible
role for X-linked genes in amygdala development. Neuropsychology
17, 39 – 49.
Ledoux, J.E., 1995. Emotion: clues from the brain. Annu. Rev. Psychol. 46,
209 – 235.
Macrae, C.N., Alnwick, K.A., Milne, A.B., Schloerscheidt, A.M., 2002.
Person perception across the menstrual cycle: hormonal influences on
social cognitive functioning. Psychol. Sci. 13 (6), 532 – 536.
Mans, M.S., Macmillan, I., Scott, J., Young, A.H., 1999. Mood, neuropsycholgical function and cognitions in premenstrual dysphoric
disorder. Psychol. Med. 29, 727 – 733.
McEwen, B.S., Alves, S.E., Bulluch, K., Weiland, N.J., 1997. Ovarian
steroids and the brain: implications for cognition and aging. Neurology
48, 8 – 15.
Morris, J.S., Frith, C.D., Perrett, D.I., Rowland, D., Young, A.W.,
Calder, A.J., Dolan, R.J., 1996. A differential neural response in the
human amygdala to fearful and happy facial expressions. Nature 383,
812 – 815.
Morris, J.S., Ohman, A., Dolan, R.J., 1998. Conscious and unconscious
emotional learning in the human amygdala. Nature 393, 467 – 470.
Osterlund, J.K., Hurd, Y.L., 2001. Estrogen receptors in the human
forebrain and the relation to neuropsychiatric disorders. Prog. Neurobiol. 64, 251 – 267.
Penton-Voak, I.S., Perrett, D.I., 2000. Female preferences for male faces
change cyclically. Further evidence. Evol. Hum. Behav. 21, 39 – 48.
Sanders, G., Sjodin, M., Chastelaine, M., 2002. On the elusive nature of sex
differences in cognition: hormonal influences contributing to within sex
variation. Arch. Sex. Behav. 31, 145 – 152.
Skuse, D., 2003. Fear recognition and the neural basis of social cognition.
Child Adolesc. Ment. Health 8, 50 – 60.
Smith, S.S., Gong, Q.H., Hsu, F.C., Markowitz, R.S., ffrench-Mullen,
J.M., Li, X., 1998. GABA(A) receptor alpha4 subunit suppression
prevents withdrawal properties of an endogenous steroid. Nature 392,
926 – 930.
Tomasello, M., 1999. The Cultural Origins of Human Cognition. Harvard
Univ. Press, Cambridge, MA.
Woolley, C.S., Gould, E., Frankurt, M., McEwen, B.S., 1990. Naturallyoccurring fluctuation in dendritic spine density on adult hippocampal
pyramidal neurons. Neuroscience 10, 4035 – 4039.