‘You’re in such a bad mood. Is it that time of the month
again?’ Biologically healthy females in their reproductive years are likely
to be confronted with such statements at ‘that time of the month’,
the beginning of the period and the start of a new menstrual cycle.
The earliest records reveal that the menstrual cycle affects mood and behavior in
women and thus the central operation system, the brain. Plato suspected that the
mourning womb, grieving over not carrying a child, might cause the monthly
experienced symptoms [1]; the ancient
Greeks named the uterus after a behavioral trait probably caused by this organ,
which was only discovered in women – hystera. These popular expressions are
not unsubstantiated. Scientific studies proved that some women do experience an
increase in anxiety around their menstruation [2] and the menstrual cycle; the intake of ‘the pill’,
also affects the structure and function of the brain (for a general overview, please
see, e. g. [3]
[4]
[5]). However, precise theories of how menstrual cycle phases and the
intake of hormonal contraceptives influence and shape the human brain is yet
limited. It is precisely this question that neuroscientists have been dealing with
for more than 20 years.
To better understand the effect of sex hormones, particularly the key players of the
menstrual cycle, estradiol and progesterone, on the brain, one must first understand
their underlying modes of action, beginning with the binding to their specific
receptors. One must distinguish between four naturally biosynthesized estrogens in
women, focusing on estrogens. Estrone (E1) is predominant during menopause and is
the weakest in effect, whereas estradiol (E2) shows the most potent effects and is
present during the menstrual cycle before pregnancy and menopause. Estriol (E3) is
the predominant estrogen during pregnancy, and estetrol (E4) is also only present
during pregnancy [6]. Estrogens exert
their effects after binding to one of two to date known distinct intracellular
estrogen receptors (ER): estrogen receptor alpha (ERα) and estrogen receptor
beta (ERβ). ERs are steroid receptors and act as ligand activation
transcription factors, resulting in the modulation of gene transcription [7]. ERα and ERβ are vastly
distributed in the brain, whereas both receptors have overlapping expression
patterns in most parts of the human brain. Both receptors are present in the
cerebral cortex but with different concentration patterns in distinct cortical
layers. Whereas ERα expression dominates in the hypothalamus and the
amygdala with only a low accumulation of ERβ, the opposite applies to the
entorhinal cortex, thalamus, and hippocampal regions, which are one of the most
abundant ERβ expressing areas [8]. This distinct expression of both receptors assigns particular roles to
both receptor types, with ERα being involved in the modulation of neuronal
populations with autonomic and reproductive neuroendocrine functions, emotional
processing, affective and motivational behaviors, and ERβ modulating
cognition, non-emotional memory, and motor functions. Additionally, estrogens,
particularly estradiol, provide neuroprotective effects in the central nervous
system due to attenuation of neuroinflammation and neurodegeneration [6].
Progesterone (P4) is another sex hormone associated with neuroprotective
effects, besides its well-studied role in regulating reproduction and female sexual
behaviors. Additionally, in its role as a neurosteroid, it is involved in
neuroplasticity [9], neurogenesis [10], and neuroinflammation [11]. Because of its diverse effects, it is
not surprising that progesterone receptors are broadly expressed throughout the
brain.
The role of progesterone in the brain regarding cognitive brain functions was already
discussed 20 years ago. Hausmann and Güntürkün (2000)
described the effect of progesterone on brain lateralization of cognitive functions
and postulated the progesterone mediated interhemispheric decoupling hypothesis
[12].
The general concept of the lateralization of cognitive function to either one of the
two hemispheres is a basic principle of the organization of the human brain.
Different cognitive functions, e. g., language, spatial attention, face
processing, and memory, are distributed differently across the brain’s two
hemispheres, known as hemispheric specialization. With the advent of the development
of modern brain imaging techniques, in particular functional transcranial Doppler
sonography (fTCD) and fMRI, the study of the hemispheres of the brain is more widely
feasible and has enabled non-invasive studies addressing the issue of the
hemispheric specialization of cognitive functions in large cohorts of healthy
participants as well as patients [13].
The current method of choice for measuring brain function is fMRI, as it is
non-invasive and without any known side effects.
With the advent of these improvements in methodology, researchers could show that
various cognitive functions are mainly located in one hemisphere; for example, in
most individuals, language functions are lateralized to the left hemisphere [14]
[15]
[16]. In contrast,
visuospatial functions are processed by the right hemisphere [17]
[18]. However, these studies also highlighted that for all these
processes, the degree of lateralization is variable not only between subjects but
also within subjects. For example, language, as a typically left-hemispheric
localized cognitive function, an atypical right-hemispheric or bilateral form of
language lateralization has been observed in up to 10% of the human
population [14]
[19]
[20]
[21]
[22]. Visuospatial attention function, a
predominantly right-hemispheric cognitive function, is again subject to marked
variability across subjects [17]
[23]
[24]
[25].
Besides the hemispheric specialization of various cognitive tasks, these functions,
especially language and visuospatial processes, have also been shown to be
lateralized to a sex-specific manner. These studies found more pronounced functional
cerebral asymmetries (FCAs) in men than women [26]
[27] and consequently
highlight the role of sex hormones for inter- and intraindividual variations in
FCAs. In general, FCAs are a simple model to investigate functional connectivity in
the brain, especially between the left and right cerebral hemispheres, referring to
the relative differences in many neural functions and cognitive processes [28]
[29]
[30]
[31]. FCAs tend to be stable and more
robust in men, whereas they greatly vary in women with an overall more symmetrical
or bilateral pattern [12]
[26]. However, this is not the case over
women’s entire lifespan. After menopause and during menses, FCAs are
comparable to those in men, highlighting the role of gonadal hormones, especially
progesterone, in modulating lateralization patterns, leading us back to the
progesterone-mediated interhemispheric decoupling hypothesis. Hausmann and
Güntürkün (2000) examined the effect of gonadal hormones on
FCAs and concluded that FCAs seem to be hormonally modulated by a global mechanism:
In general, both hemispheres work as partially independent systems, with each
processing stimuli simultaneously. Such simultaneous and independent processing
requires control mechanisms to coordinate and control the outputs from both
hemispheres [12]. One coordinating key
mechanism is the interhemispheric inhibition across the corpus callosum, which
determines FCAs [32]. The corpus callosum
consists of large parts of excitatory glutamatergic pyramidal neurons’
fibers and only a small amount of inhibitory gamma-Aminobutyric acid-ergig
(GABAergig) fibers [33]. Although the
amount might be smaller, the longer-lasting effects of callosal activation are
inhibitory and can be induced pharmacologically, resulting in an attenuation of non-
N-methyl-D-aspartate receptor (non-NMDA) glutamate receptors and a
reduction of short excitatory and longer-lasting inhibitory influence [34].
The same effects are revealed by physiological doses of progesterone [35]. Thus, as they occur naturally during
the luteal phase of the menstrual cycle, high progesterone levels can reduce
transcallosal inhibition and thus lead to a functional decoupling of the hemispheres
and hence a temporary reduction of FCAs. In summary, this leads to an overall
functional hemispheric decoupling and, thus, to a temporal decrease in functional
asymmetry. The authors conclude that steroid fluctuations during the menstrual cycle
modify cerebral asymmetries to a certain extent, with the decrease in sex hormones
(during menses and after menopause), stabilizing cerebral asymmetries, and an
increase (during the midluteal phase), leading to reduced lateralization. Even more
interesting, during low hormonal phases (menstruation), female asymmetries are
similar to that of men and post-menopausal women.
Whereas the described hypothesis is 20 years old, it has received some empirical
support from different studies with various designs and techniques [28]
[29]
[36].
For example, Pletzer and colleagues compared behavioral performance using a Navon
figure paradigm. They investigated men, naturally cycling women with women during
the follicular and during the luteal phase, and users of oral contraceptives (OCs)
during the active pill phase [37]. During
the focused attention condition, luteal women showed reduced global advantage
displayed by faster responses to global vs. local targets compared to men,
follicular women, and OC users. This is underpinned by sex hormone concentration as
a global advantage during the focused attention condition related significantly
positively to testosterone levels and significantly negatively to progesterone, but
not estradiol levels. Further, interference was significantly enhanced in OC users
as compared to women with a menstrual cycle and related positively to testosterone
levels in all naturally cycling women and men. Highly interestingly, when each group
was separately analyzed, the relationship of testosterone to global advantage and
interference was reversed in women during their luteal phase as opposed to men and
women during their follicular phase. These results support the hypothesis of
progesterone-mediated inter-hemispheric decoupling, as global processing is
lateralized to the right and local processing to the left hemisphere. Additionally,
the obtained effects might result from a testosterone-mediated enhancement of
right-hemispheric functioning.
In a combined behavioral and MRI study, behavioral results and MRI activation
patterns were compared in naturally cycling women, women taking OCs, and men.
Subjects performed two distinct numerical tasks. They reported that OC-users
resemble follicular women in their behavioral performance but show male-like brain
activation patterns during both tasks [38].
A more recent fTCD study underlined the results and reported lower test-retest
reliability in women taking oral contraceptives (OC) compared to men investigating
language dominance. Interestingly, the included women showed a significant shift
from left hemisphere dominance towards bilaterality around menstruation with a
significant reversal afterward. Authors declared the menstrual cycle a source of
inconsistency and a challenge for language dominance assessment in epilepsy [39].
A newly developing research area is the effect of menstrual cycle phases on the
resting state [40]. In addition to
task-related functional brain imaging studies, these studies revealed several
functionally relevant cortical networks that exhibit synchronous fluctuations in
brain activity while participants are at rest without performing a specific task.
Resting-state fMRI (RS fMRI) has identified specific networks that are spatially
comparable to task-related activations, for example, the default mode network (DMN),
which is comprised of the dorsal and ventral medial prefrontal cortex (mPFC), the
posterior cingulate cortex (PCC)/precuneus and lateral parietal cortex [41]
[42]
[43]. Whereas the function
of this network was hypothesized to be stimulus-independent, reflecting the brain
activity, e. g., during daydreaming or mind-wandering [44], more recently, it has been suggested
that RS activity in this network reflects spontaneous, intrinsic brain activity
[45].
To date, only a few studies have investigated sex hormonal effects on RS connectivity
displaying a significant heterogeneity in terms of methodology and obtained results.
Petersen et al. (2014) applied a between-subjects design to investigate RS
connectivity in the anterior part of the DMN under different hormonal states, both
across the menstrual cycle in normally cycling women and in oral contraceptive pill
users. This study reported increased RS connectivity between the right anterior
cingulate cortex (ACC) and the executive control network and reduced RS connectivity
between the left angular gyrus and the anterior DMN during the luteal compared to
the menstrual phase [46]. However,
progesterone levels were unusually high during the menstrual phase, resulting in
only a small difference in hormone concentration during the luteal phase.
Additionally, no cycle-dependent estradiol differences were obtained; thus, the
included subject might have been inaccurate in their cycle phase self-reports.
Hjelmervik et al. (2014) investigated four fronto-parietal (cognitive control) RS
networks in a repeated measures design and could not find any cycle-related effect
on RS connectivity [47]. These results
are in line with those obtained by De Bondt et al. (2015). They also did not find
any effect of sex hormones in fronto-parietal networks. However, in the DMN, an
increase in RS connectivity between the network and the cuneus was observed in the
luteal phase compared to the follicular phase [48].
Arélin et al. (2015) conducted 32 RS scans on a single subject across four
menstrual cycles. Initial analyses revealed that high progesterone levels were
associated with increased connectivity of the dorsolateral prefrontal cortex (dlPFC)
and the sensorimotor cortex to the RS network. A region-of-interest analysis
revealed that high progesterone levels were associated with higher RS connectivity
between the right dlPFC, bilateral sensorimotor cortex, the hippocampus, and the
left dlPFC and bilateral hippocampi during rest [49].
Finally, Weis 2019 et al. investigated sex hormonal effects on RS connectivity in the
DMN and described variations in RS connectivity across the menstrual cycle [43].
Studying women during different cycle phases, as the menstrual cycle serves the most
dramatic hormonal changes within short periods, has become a significant tool to
investigate the influence of sex hormones on cognitive behavior and its underlying
functional brain organization.
This is from potential interest, as different menstrual cycle phases also affect
cognitive function and psychological well-being.
Previous behavioral and neuroimaging studies have particularly shown the effects of
the menstrual cycle on cognitive functions, for example, visuospatial ability [36], verbal skills [50]
[51], and emotional memory [52]. Focusing on psychological well-being, a recently published meta-analysis
summarizes that the menstrual phase during the menstrual cycle is associated with a
greater risk of serious mental health outcomes, e. g., suicide attempts,
psychiatric admissions, and drug abuse [53].
Thus, understanding the underlying principles of the mode of action of sex hormones
in the brain is of high interest in the scientific community.
The role of ‘the pill’
To make it even more complex, women experience dramatic hormonal changes during
their lifetimes. With the advent of puberty, sex hormone levels increase
immensely, with mean values being comparatively high for the following more than
30 years, known as reproductive years [54] before they relatively abruptly decline during the transition to
reproductive senescence, commonly known as menopause [55].
However, women do not exclusively experience substantial alterations of sex
hormones during their lifetime, but, with a focus on the reproductive years, in
an approximately monthly manner, known as the menstrual cycle, but also due to
the intake of OCs. Worldwide, more than 100 million people use hormonal
contraceptives in the form of oral contraceptives (OCs) as a method of choice
for contraception [56]; in Germany,
it is more than half of the female population in their childbearing years
between 18 and 49 years [57].
Comparable to the number of existing neuroimaging studies investigating the
influence of naturally occurring sex hormones on brain functions is already
limited, the number of studies examining the effect of synthetic hormones,
particularly ‘the pill’, on the brain is also low. Surprisingly
little is known about how the OCs affect the brain function of the user. For a
decade, OCs were acknowledged to show an altered mate preference compared to
non-OCs users [58] and different
brain activation patterns while watching erotic stimuli [59]. Whereas these studies shed the
first light on this topic, they are, of course, also entertaining and not only
interesting to professional scientists. However, systematical investigations of
OCs dependent changes in ‘classical’ robustly lateralized brain
functions, e. g., language or visuospatial attention, using fMRI, are
still rare. Rumberg and colleagues (2010) showed increased activation in
right-hemispheric task-specific areas in OC users compared to non-users during a
word generation task [51]. In
addition, Pletzer and colleagues (2014) found more lateralized brain activation
patterns in numerical tasks [38], in
which cognitive demands can be related to spatial abilities [60]. Further brain function
differences between OC users and non-users are described during resting state
[46], and, for example, for
reward- and face processing [61]
[62].
The influence of the menstrual cycle and ‘the pill’ on brain
structure
Brain structure differences between men and women have been described in several
studies [63]
[64]
[65], with larger brains in males than
in females, on average; for example, larger gray matter (GM) volumes in
amygdalae, hippocampi, and temporal pole and orbitofrontal gyri in men, whereas
women show larger thalami, precuneus, right insula cortex and right anterior
cingulate gyrus [64]. Focusing on
diffusion tensor imaging (DTI), studies comparing male and female brain anatomy
are still rare. Menzler et al. (2010) described regional microstructural
differences between male and female brains within the thalamus, corpus callosum,
and cingulum. They concluded that higher values of fractional anisotropy and
lower radial diffusivity in these areas were caused by differences in
myelination between men and women [66]. These results are in line with the results of Dunst et al.
(2014), who also described differences between male and female brains with
regard to myelination [67].
Still, it’s challenging to compare male and female brains, as the
potential influence of sex hormones is hard to detangle. Women experience strong
hormonal changes in sex hormones over their lifespan, whereas these levels are
relatively constant in men. Therefore, investigating the cyclic fluctuation of
female sex hormones during the menstrual cycle is a suitable tool to study these
effects within a short period.
The fluctuation of female sex hormones during the menstrual cycle does affect not
only brain function but also brain structure. The first results on this topic
were reported over ten years ago by Protopopescu et al. (2008), who compared
women in their late follicular phase (high estradiol, low progesterone) and
mid-luteal phase (medium estradiol, high progesterone). They found increased
gray matter (GM) volumes in the right anterior hippocampus and decreased values
in the right globus pallidus and putamen in the late follicular phase [68]. These results were confirmed by
Lisofsky and colleagues (2015) and, additionally, in a longitudinal single
subject study [69]
[70].
Partly supporting results were described by Pletzer et al. (2010), who described
slightly larger GM volumes in the right parahippocampal/fusiform gyrus
during their early follicular phase (low estradiol and progesterone) compared to
their midluteal phase (medium estradiol and high progesterone levels) [71]. Further described brain regions,
which are affected by different menstrual cycle phases, are the right middle
frontal gyrus (MFG), the right anterior cingulate cortex (ACC), and the left
insula. These regions showed larger volumes during the pre-ovulatory phase
compared to the midluteal phase [72]
[73]. Increased volumes
in the left MFG but opposite results for the ACC were reported by Protopopescu
et al. (2008) [68]. In a recent study
on a sample of 55 women to assess menstrual cycle-dependent effects, there was a
significant pre-ovulatory estradiol-driven increase in bilateral hippocampal GM
volumes and a significant progesterone-dependent increase in GM volumes of the
right basal ganglia in the mid-luteal phase [74]. Further information concerning
overall structural changes across different hormonal states within a
woman’s life, including brain maturation, puberty, menstrual cycle, OC
intake, pregnancy, and menopause, is available within a recently published
systemic review by Rehbein and colleagues (2020) [3].
Summarizing the above-described results, the hippocampus, basal ganglia, and
insula are possible targets of structural changes due to sex hormones
fluctuations during the menstrual cycle, accompanied by trend findings in
parahippocampal and fusiform regions, the ACC and MFG.
The research field on the influence of the pill on brain structure is even
younger. Around 10 years ago, Pletzer et al. (2010) published the first
exploratory study [71]. They reported
larger GM volumes in the prefrontal cortex, ACC, parahippocampal and fusiform
gyri, and cerebellum in women using OCs. However, neither the
‘pill’s generation’ nor the chemical combination of the
OC was considered. That these issues hold a strong influence regarding brain
structure, particularly GM volumes, could already be confirmed in a follow-up
study conducted by the same authors a few years later. Pletzer and colleagues
(2015) investigated the consequences of the use of so-called androgenic and
anti-androgenic OCs, referring to their receptor binding properties and thus
their ability to stimulate male characteristics [75], resulting in opposed effects on
brain structure [76]. Whereas
anti-androgenic OCs lead to larger gray matter volumes compared to women with a
natural cycle, users of androgenic OCs displayed partly smaller brain regions in
specific brain areas. In particular, whereas antiandrogenic OCs lead to larger
GM volumes compared to women with a natural cycle in bilateral fusiform gyri,
the fusiform face area (FFA), parahippocampal place area (PPA) and the
cerebellum, users of androgenic OCs displayed significantly smaller brain
regions in the bilateral middle and superior frontal gyri.
However, the authors did not control the exact hormone derivatives in the
combined preparations. In general, the concentration of progesterone and
estradiol derivates has been gradually reduced over the last decades to reduce
side-effects [77]. However, different
types of combinations may also still be associated with different side effects
[78]. Whereas some progesterone
derivates are considered to have androgenic properties, others, such as
drospirenone and desorgestrel, may show anti-androgenic effects on the brain
[79]
[80]. The latter ones have also been
postulated to be favorable regarding mood symptoms [81]. Besides the oral intake of
hormonal contraceptives, alternative administration routes have been developed.
Thus, hormones can be administered vaginally or transdermal. Additionally,
long-acting-reversible contraception such as injections, implantable devices,
and progesterone releasing intrauterine devices are effective contraceptive
options [82]. However, the effect of
these hormone administration pathways on the brain has not been extensively
studied. This includes levonorgestrel-intrauterine-devices, although they are
one of the most used contraceptive methods worldwide. Our findings align with a
recently published study by Bürger et al. (2021) [83], who could not find these studies
either. The lack of MRI studies investigating the effect of intrauterine devices
on the brain might be their incompatibility with MRI scanners. There is a risk
that the IUD may slip, reducing its contraceptive effect.
Gender identity and brain structure
A recently published study showed that even the identification of a gender role
affected grey matter volume [84]. The
author corroborated findings of sex hormones on brain structure and demonstrated
testosterone-driven effects in women to more male-like brain morphologies.
Furthermore, estradiol led to more female-like brain morphologies. The author
described a positive association between a more feminine gender role and a more
female-like brain morphology in men, notably concerning the left middle frontal
gyrus. Additionally, differences in gender roles and gray matter volumes between
OC-users and NC women were described [84]. Interestingly, focusing on the left middle frontal gyrus, this
brain region is typically larger in women and has already been addressed in an
earlier study where researchers reported larger cortical thickness in untreated
male-to-female transsexuals compared to men [85]. These results are in line with
prior results. A prior study showed that androgen treatment increases the female
brain’s volume towards male proportions, and anti-androgen and estrogen
treatment reduced the size of the male brain towards a female morphology [86]. The findings imply the plasticity
of the adult human brain structure towards the opposite sex under the influence
of cross-sex hormones [86]
[87].
Effects of gender-affirming hormone therapy (GHAT) on brain structure and
function
Focusing on gender-affirming hormone therapy (GAHT) in transgender individuals to
obtain their desired gender phenotype, longitudinal studies show that this
therapy either feminizes brain structure in Male-to-Females (MTFs) or
defeminizes brain structure in Female-to-Males (for a review please see [88]). Particularly, in MTFs, a
duration of four months of anti-androgen and estrogen GAHT resulted in decreased
brain volumes in the right hippocampal region and increased ventricle volumes
compared with male controls [89].
Additionally, the decrease in brain volume was correlated with changes in
progesterone levels. Another study reported an increase in total brain and
hypothalamic volume and decreased ventricle volumes compared with female
controls [90]. However, these studies
display plastic changes in specifically subcortical structures related to memory
and emotional processing [88].
Regarding brain function, Sommer et al. [91] detected a potential influence of
three months of GAHT on brain activation patterns during language and mental
rotation tasks in eight MTFs and six FTMs individuals with lateralization of
both evoked activations remaining stable. Additionally, they reported a
correlation between the total increase of language-related activation after GAHT
with post-treatment serum estradiol levels and post-treatment testosterone
levels with full brain activation during mental rotation. In summary, the
application of MRI to the investigation of the transgender brain and the effects
of GAHT is still in its infancy, and data is yet only derived from small sample
sizes; however, it offers an excellent tool to understand regional and network
effects of hormonal treatments on the brain.
What is the challenge of investigating the effect of sex hormones on the
brain?
Despite the early research interest in this topic in general over 80 years ago
[92], it is highly surprising
that only a handful of scientists worldwide are actively examining the effect of
sex hormones on the brain, no matter if they occur physiologically during a
woman’s lifespan (e. g., puberty, pregnancy, menopause), or how
the application of hormonal contraceptives influence and manipulate their
effects. One should think that nowadays, elaborated neuroimaging methods are
feasibly available and could easily shed light on this intriguing topic,
affecting millions of women worldwide. Thus, it might sound surprising at first
glance; however, scientists have been aware of methodological issues and
challenges since the early 70 s. Here, Sommer (1973) reviewed 33
publications investigating the effect of the menstrual cycle on cognition and
perceptual motor behavior and found no evidence. However, she concluded that
this result might be due to methodological problems, which researchers are
confronted by when investigating the menstrual cycle [93]. This issue might be an
explanation for the relatively small amount of neuroimaging studies.
Additionally, these studies, which are reviewed to a significant part by
Sundström Poromaa & Gingnell (2014), are not consistent
concerning the obtained results and the described hormonal effects [94]. Not only is it very time
consuming, but also the exact determination of hormonal state, e. g.,
from collected blood or salvia samples, has often not been state of the art yet.
For example, in Protopopescu and colleagues’ study (2008), the menstrual
cycle phase definition was very lenient, as it did not include hormone analyses
[68]. Pletzer and colleagues also
relied on verbal reports in one of their studies [38], and so did Rumberg et al. 2009
[51].
Furthermore, infrastructural issues occur; for example, MRI measurement
appointments need to be reserved and are often not spontaneously available;
blood- or salvia samples need to be pretreated and correctly stored. These
additional barriers and challenges might also explain the usually meager number
of initially investigated subjects (e. g., [71] only included 14 subjects) and the
high drop-out numbers caused by later correction of the cycle phases.
Concerning the studies investigating the effects of OC, a further difficulty
occurs: OCs are available in various combinations of synthetic hormones,
particularly with regard to their androgenic modes of actions. These issues hold
a strong influence regarding brain structure, as has been confirmed for instance
in a follow-up study by Pletzer and colleagues (2015). As written above, they
examined the effects of androgenic and antiandrogenic OCs, resulting in opposed
impacts on brain structure. However, again, they did not control for the exact
hormone derivatives in the combined preparations [95].
These methodological challenges have already been clearly recognized several
years ago. Pletzer and Kerschbaum (2014), for instance, stated almost ten years
ago that more systemic research is needed to “reveal the true nature
of OC-dependent effects on cognition as well as the impact of synthetic
steroids on neuronal correlates”
[79]. Accordingly, previous study
results must be considered with reservations, as different cycle phases were
compared, relatively small sample sizes were examined, and exact hormonal
determination is not present in all studies. Past studies relied on
self-reports, which were rather unspecific and accompanied by high drop-out
rates. Other studies did not determine hormonal concentrations at all;
consequently, data collection on the requested cycle phase cannot be guaranteed.
Focusing on the effect of OCs on brain structure, former studies did neither
control for ‘the pill generation’ nor the exact chemical
combination. Additionally, different analysis pipelines were applied, using
different brain parcellations, thus impeding the comparability of the yielded
results.
With regard on brain function, fMRI studies, the method itself is challenging.
With a boost of awareness regarding a concerning shortage of reliability and
reproducibility in neuroscientific research, the degree of validity of the
yielded results has become uncertain. Various forms of instability have been
identified in structural and functional measurements, including across operating
system versions [96], minor noise
injections ([97], and data set or
implementation of theoretically equivalent algorithms [98]
[99]. These issues hold practical
applications in order to decide which tool/implementation should be
applied for an experiment [100].
Focusing on conventional fMRI, regional brain activity is estimated by measuring
the BOLD signal that indicates changes in blood oxygenation associated neural
activity [101]. Commonly, researchers
map brain activity evoked by specific cognitive functions by contrasting the
regional BOLD signal during a control condition with the BOLD signal during a
condition of interest [102]. Thanks
to this approach, task-fMRI enables unique insights into the brain, ranging from
basic perception to complex thought and, with a clinical focus, the opportunity
to directly measure neurological and psychiatric dysfunction [102]. The original idea of task-fMRI
was to examine functions of the average human brain by measuring within-subject
differences in brain activation between task and control conditions and
averaging them together across subjects to obtain a group effect, resulting in
mostly robust brain activity. This led to the idea of using the same paradigms
to study between-subject differences. Thus nowadays, fMRI is widely used for
studying how the brains of individuals differ. However, the reliability of the
most commonly applied paradigm is largely unknown and an object of current
debate within this research field [103]
[104]
[105]. Recently, concerns have been
raised that the conclusions drawn from some neuroimaging studies are either
bogus or not generalizable. This might be caused by the high vulnerability of
fMRI results to low statistical power, flexibility in data analysis, software
error, and a lack of direct replication [106].
Focusing on visuospatial attention, by now, available paradigms investigating
visuospatial functions provide widely distributed activation patterns and
markedly inter- as well as intra-subject heterogeneity; thus they would add
further variability and might reduce the obtained effect sizes [107].
Behavioral studies show that spatial tasks favor men and that women during
high-hormonal phases, notably late follicular- and luteal phase, score lower on
mental rotation tasks than during the low-hormonal phase [108]. Surprisingly, the effect of OCs
and particularly their androgenic activity has not been systematically
investigated using fMRI, despite a behavioral study impressively demonstrated
that performance in the applied mental rotation task was best in OC users on an
androgenic treatment compared to users of antiandrogenic OCs and nonusers [109].
Necessary prerequisites to study the influence of sex hormone on the brain
using MRI
These limitations, however, highlight that future refinement of the utilized
paradigms is strongly needed and is an essential prerequisite for a more
thorough investigation of the right-hemispheric lateralization in visuospatial
attention. This is even more important before applying these paradigms,
providing rather heterogeneous activation patterns across subjects to research
questions, which expect to obtain comparatively small differences. This is the
case when studying the effects of sex hormones on the brain. Regarding data
collection and analysis, it is thus recommended to include robust and reliable
fMRI paradigms to increase the obtained data’s validity.
Future studies are advised to consider the following recommendation concerning
the study design: Investigating women with a menstrual cycle, favorably several
cycle phases is favorable and depends on the research focus. Concerning this
point, different questions concerning the effects of menstrual cycle phases and
OC intake need to consider distinct cycle phases: Whereas studying the potential
effects of sex hormones on different brain functions, particularly their degree
of lateralization, might be most suitable for comparing high- and low
progesterone driven cycle phases (with regard to the progesterone dependent
hemispheric decoupling hypothesis [12]; structural differences might be more prominent during high- and
low estradiol differing phases [68]
[69]
[71]
[72]
[73].
Still, it is not possible to narrow down the obtained results to either estradiol
or progesterone as both hormones are constantly present in the organism and
interact with each other. The endogenous release of progesterone during the
luteal phase is always accompanied by release of estradiol. The release of other
hormones could also alter the responsiveness of progesterone or estradiol in the
brain [110]
[111]. However, the use of
sophisticated statistical analyses, for example linear mixed effects models, has
made it possible to identify hormonal key players within a specific cycle phase.
This procedure is well described in Pletzer et al. 2018 [74]. Additionally, it is important to
validate cycle phases by evaluating the exact hormone concentration in,
e. g., blood samples. When investigating women under an OC-treatment, it
is highly recommended to include only one explicit OC-type per test group,
containing the exact amount of estradiol and progestin derivatives. For all
women, a sophisticated anamnestic interview is advised, including each
individual’s hormonal history, e. g., previous pregnancies, use
of other oral contraceptive types, or hormonal replacement therapies, as these
have shown an impact on the human brain [112]
[113].
Furthermore, scientists have to be careful with regard to the results’
interpretation. As menstrual cycle phases are governed largely by the
concentration and fluctuation of the captured female sex hormones estradiol and
progesterone, these might not be the only potential important influence factors
affecting GM volumes. As already described earlier, the natural menstrual cycle
as well as OCs affect various metabolic processes (e. g., basal body
temperature, heart rate, and breathing patterns), which might affect the
measured MRI signal.
Additionally, a recently published study described that menstrual cycle affects
cerebral blood flow CBF as well, which is in turn crucial for functional MRI,
determined by changes in the BOLD signal [114]. It is already known that estradiol has excitatory and
vasodilatory effects in arteries, which could lead to widespread increases in
CBF. In contrast, progesterone may have opposing effects on CBF. As estradiol
and progesterone receptor density vary across the cortex [115], thus effects may be stronger in
specific brain areas. To elucidate the potential effects of hormones on blood
flow on the brain, Cote et al. directly investigated the link between CBF and
estradiol and progesterone, while controlling for the size of the large feeding
arteries using a multi-modal approach combining arterial spin labeling (ASL) and
non-contrast-enhanced Time-Of-Flight (TOF) magnetic resonance angiography [114]. They observed a relatively
strong, inverse relationship between progesterone levels during the luteal phase
and CBF during the same phase, with the strongest link in frontal cortex. Serum
estradiol during the follicular phase tended to correlate weakly with CBF during
this cycle phase. Additionally, during the luteal phase, estradiol did not
impact the relationship between progesterone and CBF, nor did the lumen
diameters of the large arteries feeding the anterior and posterior circulation.
They concluded that estradiol and progesterone have strikingly different and
independent effects on CBF, which are unlikely to be driven by large artery
morphology. Furthermore, they showed that CBF is dynamic and related to the
hormonal state in women. These results are crucial with regard to fMRI studies,
as these results are obtained from the BOLD signal, which in turn is affected by
the CBF.
Therefore, the results cannot be interpreted solely as a direct effect of
fluctuating sex hormones on the brain but could also be evoked by further
physiological parameter changes. Additionally, the decision to start or end
treatment of OCs might also be accompanied by changes in personal circumstances
that, in turn, may affect overall psychological well-being.
Thus, it is practically impossible to detangle and identify the sole effect of
the hormonal key players of the menstrual cycle on the female human brain.
Additionally, fMRI paradigms that assess emotion and empathy processing, as well
as social interactions, should be included to investigate sex hormones’
effects on these essential interpersonal functions. Including various
neuropsychological tests and questionnaires could further increase knowledge
about the hormonal effects on cognition and psychological well-being.
An elucidation of how hormones affect our brain will also help us to better
understand disorders such as premenstrual syndrome, and postnatal depression and
ultimately to be able to treat them successfully. This may also help us to
understand what might have motivated ancient Greek anatomists to name the uterus
after a former psychiatric illness.
