Mapping changes in women's visual functions during the menstrual cycle: narrative review

ABSTRACT BACKGROUND: This article systematically updates the literature on changes in visual functions during the phases of the normal menstrual cycle in women. OBJECTIVES: To update Guttridge's 1994 review of visual structures and functions associated with the menstrual cycle and broaden the search through psychophysical, neuroimaging and neurobehavioral measurements covering 1994-2020. DESIGN AND SETTING: Narrative review conducted in a neurosciences and behavior laboratory in Brazil. METHODS: The PubMed, Cochrane Clinical Answers and Google Scholar databases were searched. After screening and applying the eligibility criteria, 32 articles were examined. Through this analysis, the following information was extracted: (1) geographical distribution of the study; (2) sample size (according to age and phase of the menstrual cycle); (3) type of measurements according to psychophysical, neuroimaging and neurobehavioral instruments; (4) vision testing model; (5) visual subcategory evaluated; (6) categories of processed visual stimuli; and (7) main findings. RESULTS: The menstrual phases give rise to significant changes in visual functions, including in relation to orientation and spatial attention, visual campimetry and visual sensitivity. These relate specifically to the follicular and luteal phases. CONCLUSIONS: These findings theoretically expand the effects of menstrual cycles on visual functions found by Guttridge (1994). Despite some inconsistencies in the studies analyzed, it was found that visual processing during the follicular and luteal phases of the normal menstrual cycle of healthy women can explain physiological, cognitive, behavioral and social modulations.

menstrual cycle. 6 The main findings from that review pointed to variations in thickness, curvature, corneal sensitivity, intraocular pressure, maturation of the connective membrane and production and secretion of vicarious hemorrhagic tears in the eye. Among the functional measurements, changes to flashlight thresholds, detection and discrimination of visual stimuli, visual acuity and color vision can be highlighted. 7 Most of the studies discussed had an ophthalmological or clinical focus and did not explore the psychophysical, neurobehavioral and electrophysiological functioning of vision. 6,8 Although several conclusions about ophthalmic function performance during the phases of the normal menstrual cycle were reached in previous studies, these data presented some conflicting features, such as with regard to increased visual sensitivity in the premenstrual period. 6 Thus, there is a need to update the data sources on visual function performance during the menstrual cycle.

OBJECTIVES
The phases of the menstrual cycle may be clinically significant in relation to the mechanisms of visual processing, as this may facilitate understanding of the mechanisms underlying visual functions. To this end, the aim of this study was to investigate the methodological practices involved in studies on the normal menstrual cycle, regarding processing of the visual system of healthy women and the underlying visual mechanisms.

METHODS
We carried out an up-to-date search for studies that describe the use of psychophysical, neuroimaging and neurobehavioral tools for analyzing the visual functions of women during the normal menstrual cycle. Our search covered the period from the Guttridge study conducted in 1994, up to 2020. The following databases were used: PubMed, Cochrane Clinical Answers and Google Scholar. The following terms were used: "visual functions" OR "visual system" OR "visual perception" OR "vision" AND "menstrual cycle" OR "estrous cycle" OR "menstrual period" OR "ovarian cycle". The reference lists of the studies found were also reviewed to identify any additional studies. This systematic review was reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The keywords were chosen even in the absence of the specific Medical Subject Heading (MeSH) term, in order to prioritize sensitivity over specificity.

Research strategy
During the first screening, two reviewers (BG and MT) evaluated the titles and abstracts of each study and excluded those that did not meet the eligibility criteria highlighted below. For each study selected, the full article was read and evaluated to verify whether it met the criteria that had been established. If any disagreements arose, a third evaluator (MA) was contacted. At the screening stage, all titles were analyzed, and studies that were considered important for the topic addressed were selected. Then, the abstracts corresponding to the titles that had been selected were read. Studies that were considered irrelevant to the objective of the present study were excluded. Lastly, the articles thus selected were read in full and further exclusions were made in accordance with the eligibility criteria.

Eligibility criteria
The inclusion criteria were that the studies need to have the following characteristics: (1) randomized clinical trials, quasi-randomized clinical trials, case-control or cohort studies; (2)  (4) studies that were carried out during pregnancy or menopause; (5) studies exclusively on use of contraceptives; and (6) letters, editorials, systematic reviews and bibliographic reviews. Studies that showed insufficient information, such as in the statistical analysis, or that showed incomplete procedures, were also excluded.

Quality assessment
The articles were evaluated based on internal validity (e.g. selection bias, performance bias, friction measurement bias and reports) and the validity construct (e.g. suitability of the operational criteria used). In general, the quality of the evidence from the studies was assessed using three main measurements: (1) limitations (poor design, for example); (2) consistency of results; and (3) accuracy (ability to generalize findings and provide sufficient data). Studies that failed with regard to these points were not added or selected.

Data extraction
After assessing the articles in accordance with the pre-established criteria, the following information was extracted for analysis purposes: (1) geographical distribution of the study; (2) sample size (according to age and phase of the menstrual cycle); (3) type of measurements according to the psychophysical, neuroimaging and neurobehavioral instruments; (4) vision testing model; (5) visual subcategory evaluated; (6) categories of processed visual stimuli; and (7) main findings. "NS" (not specified) was used in the tables to denote situations of insufficient information from the studies to determine the data, according to the category. Figure 1 shows a detailed flow chart of the study selection process. The tabulation showed that a total of 320 potentially eligible articles were found. Many articles were found to be incompatible with the purpose of the present study and/or featured women with irregular menstrual cycles, women using contraceptives, pregnant and lactating women, or women entering menopause.

RESULTS
In addition, some studies did not directly assess measurements of visual function during the different phases of the normal menstrual cycle. Thus, after screening and applying the eligibility criteria, a total of 32 articles were examined. Table 1 summarizes the articles according to the psychophysical methodological criteria. A total of 12 articles were found to fit within the criteria that had been established. Table 2 summarizes the articles according to the neuroimaging methodological criteria (studies with spatial and temporal paradigms). A total of six articles were found to fit within the criteria that had been established. Table 3 summarizes the articles according to the neurobehavioral methodological criteria (studies with cognitive and behavioral paradigms). A total of 14 articles were found to fit within the criteria that had been established.
The planning of comparisons, synthesis and interpretation of results were grouped thematically in the discussion, as described below.

DISCUSSION
Perceptual measurements involve basic and cognitive visual functions that describe how the visual system recognizes and interprets elementary sensory stimuli and/or complex visual scenes. 9 In this way, visual tests make it possible to investigate how vision detects, processes and recognizes the presence of light and the shape, size and color of visual stimuli, and enable investigation of visual acuity, visual field, evidence of monocular and binocular vision and even image quality. 9 For visual neuroscience, it is important to describe and understand how these measurements fluctuate with the normal menstrual cycle, i.e. how the peaks and endogenous levels of estradiol and progesterone are related to functional and perceptual changes in vision. 10 A study of this nature is important even considering that parts of this research are usually conducted with samples of men and women together, which makes it difficult to interpret the results.

Sample parameters
Most studies have had randomized blinded repeated-measurement research designs with small numbers of samples.    However, the research has varied according to the theoretical paradigms and methodological procedure used. For example, studies with psychophysical methods had at least 8 and a maximum of 32 participants, studies with electrophysiology and neuroimaging (for example, visual evoked potential (VEP) and functional magnetic resonance imaging (fMRI)) had samples with a minimum of 11 and a maximum of 30 participants. [11][12][13][14] In studies with neurobehavioral paradigms, the sample size was found to be a minimum of 8 and a maximum number of 202 women. 15 Even if the sensory threshold is a probabilistic value, its momentary values are normally distributed, except for the sample errors.
However, it is important to highlight that an approach based on sampling principles can underestimate hypothesis errors and cross-validation of studies. Therefore, solutions for increasing the sample size should be investigated, so as to address the possibility of increasing the data homogeneity. 16

The normal menstrual cycle
The menstrual cycle shows cyclical variations in the levels of various sex hormones. The effects of these hormones are not limited just to the reproductive system but affect other systems too, including the nervous system. All the participants included in the studies had a normal menstrual cycle, but comparative studies between women with a normal menstrual cycle and women who had premenstrual syndrome were also found, along with comparisons with men over periods of time similar to the menstrual cycle. [17][18][19][20] As mentioned earlier, the menstrual cycle can be divided into the follicular, ovulatory and luteal phases. 21  The results from simulations and experiments involving women were reviewed to assess the usefulness of these adaptive procedures and the special circumstances in which one might be superior to another. 45   cycle. Also, from a physiological perspective, this means that these changes have anatomical-physiological connections, with visual pathways integrated with endogenous hormones.

The neuroimaging method
Functional neuroimaging studies involve experimental baseline conditions that reveal time-space models of visual function. A 3T scanner was used in one study to assess the nature of amygdala activation using a T2 relaxation protocol and the Vienna Emotion Recognition Tasks (VERT-K) was used in another study to verify recognition of facial emotions. In both of these studies, the aim was to understand cognitive functions relating to perception of visual stimuli of emotions, in the follicular and luteal phase. 35 The recognition of facial emotions was also analyzed by means of T1e-weighted images and diffusion (DTI). 25 Other models of visual temporal processing were used to measure the latency and amplitude of electroencephalographic components in occipital areas. 14,37,41 VEP and fMRI are techniques that can be used to understand visual neural events during the normal menstrual cycle. 46 Together, all these research studies provide temporal and spatial mapping of brain function. On the other hand, all of the studies that evaluated visual functions had the problem of analyzing results that corresponded to measurement of the response to a stimulus after it had been processed by the central nervous system. The response was therefore the sum of the structural visual function and the constructive mental information.

The neurobehavioral method
This basic and applied research practice uses tests, instruments and behavioral assessment protocols to describe brain functional behavior. Articles in which visual recognition behavior was measured with regard to facial attractiveness of facial stimuli and emotional expressions were found. 15,17,26,27,30,32,35,36,[38][39][40]47 These studies used a variety of stimuli: for example, prototypes of male and female facial stimuli of different ethnicities at different stages of the menstrual cycle. 39 In addition to these models, neurocognitive studies were found that evaluated visuospatial memory and visual working memory. 29

Spatial vision: monocular and binocular measurement
It is important to highlight that some studies used the participants for recording binocular and monocular measurements. 11,18,31,34 Binocular rivalry occurs when markedly different entries for the two eyes initiate alternations in the perceptual domain between the views of the two eyes. 11 In psychophysical studies to detect and measure visual thresholds, it has been argued that there are no differences between binocular and monocular measurements. 45 This hypothesis was confirmed by Giuffrè et al. who found no differences between the left and right eye regarding visual acuity and luminance contrast. 31  However, no significant changes or trends in menstrual phases were found with regard to the temporal dynamics of ocular dominance.
This study pointed to the need for new research perspectives that would investigate GABAergic systems in relation to the dynamics between hormonal steroids and binocular rivalry.

Visual acuity
The ability to visually detect something depends on the arrange-

Visual sensitivity
From studies measuring visual sensitivity, it has been suggested that light may be a major factor for the hormonal secretory system. However, these studies have not ascertained any central mechanism for its action. For example, in one study, it was investigated whether hormonal alterations might affect the adaptation to light of four women in the menstrual, follicular and luteal phases. 28 Variability in visual sensitivity was found at all stages, but there were no specific individual differences. In another study, the effect of light on reproductive hormones was evaluated and the role of photoreception based on melanopsin sensitive to blue (480 nanometers), for mediating the non-visual effects of light, was addressed. That study was carried out among healthy women with measurements in the menstrual, follicular and luteal phases. 43 The results showed that bright blue light gives rise to secretion of follicle-stimulating hormone in women in the middle follicular phase until the end of the menstrual cycle during the morning, which suggested that a direct functional link between light and the reproductive system existed. The data were seen to be sensitive to the capacity for visual adaptation.

Luminance contrast
Luminance contrast sensitivity to stimuli of vertical sinusoidal grids of frequencies 0.5, 1, 2, 4, 8 and 16 cycles per degree of visual angle (cpd) in the menstrual, ovular, and luteal phases was assessed in one study. 31 No significant differences were found in relation to any of the spatial frequencies tested. Similarly, the luminance contrast of Gabor sinusoidal stimuli was also measured at spatial frequencies of 1, 4 and 16 cpd, in 21 women in the menstrual, ovulatory and luteal phases and in 14 men. It was shown that the sensitivity to visual contrast did not differ according to sex or use of oral contraception. Nor did it vary due to hormonal changes over the course of the menstrual cycle. 20 It was suggested that changes to the phases of the menstrual cycle did not produce any changes in sensitivity to visual luminance contrast.

Visual color
Color discrimination was assessed using the Farnsworth-Munsell 100 test in one study. 31 The results showed that although all the women in that study could be defined as normal from an ophthalmological point of view, they showed variation in their scores, with better performance in the ovulatory phase and greater numbers of errors of disposition of the records in the menstrual and luteal phases. In addition, there were no significant differences in measurements between the left eye and the right eye. The authors argued that the differences found might be associated with cyclical physiological functions or with the women's psychological state, even if imperfectly expressed. It was concluded that variations in color perception might occur during the menstrual cycle, although the differences were subtle.

Visual field
The area of the visual field was measured monocularly in one study, using automated perimetry with a short wavelength, among 48 eyes of healthy women and 45 eyes of women with diabetes mellitus, who were in the follicular and luteal phases of the menstrual cycle. 34 To highlight the results, the authors pointed out that the achromatic visual field tests were not statistically different between diabetic women and the control group. However, the mean values of the temporal retinal region were significantly lower in the luteal phase than in the follicular phase in the diabetic women. In another study, standard achromatic perimetric analysis was performed using automated short-wavelength perimeters of monocular shape, among 59 healthy women in the follicular and luteal phases, and the results were compared with those from 54 men. Unlike in the previous study, a reduction in the average value in the automated perimeter field was found for short wavelength in the luteal phase. 18 In the above two studies, 18,34 the authors were of the opinion that the hormonal behavior that occurs in the luteal phase correlates with a reduction of loss of sensation in the visual field. However, further clinical trials relating to synchronization and visual campimetry should be conducted in order to provide more accurate information.

Spatial orientation
The perception of verticality was evaluated in one study through the binocular visual acuity of 30 women in the follicular phase and 24 women in the luteal phase, and their results were compared with those of 30 men over the same time periods. The subjects had been matched previously with regard to the menstrual cycle using a mechanical rod and frame test system. 19 A higher proportion of vertical positioning errors was found among the women in the middle luteal phase than among the men or among the women in the follicular phase, which suggested that the phase was associated with the perception of verticality. Similarly, the spatial orientation of 24 women in the menstrual and luteal phases and 14 men was assessed in another study using a visual line bisection task to investigate functional brain asymmetry. 23 It was shown that high levels of estradiol during the luteal phase were related to decreased hemispheric asymmetry of spatial attention, with greater bias towards the left hemisphere. 19 These findings contribute to the idea that the luteal phase can present strong evidence of fluctuations in orientation and spatial attention.

Eye movement
The optokinetic function of 12 women in the menstrual, follicular, half-cycle and luteal phases was measured in one study.
The authors suggested that there was no difference in the pattern of eye movement. The results from that study demonstrated that hormonal changes that occur throughout the menstrual cycle had no significant effect on the speed of fixation movements. 24

Visual electrophysiology
Several studies have investigated the electrophysiological changes in visual functions that occur during the menstrual cycle of women. For example, pattern-reversal visual evoked potentials (PRVEPs) were investigated among 30 healthy women in the menstrual, follicular, ovulatory and luteal phases. 14 Although the results were not statistically significant, there was greater mean amplitude of the P100 during the ovulatory phase in relation to all other phases. Thus, the P180 latency was statistically lower in the follicular phase than in the luteal phase. In another similar study, changes to the potential relating to social stimulus events were investigated among women in the menstrual, ovulatory and luteal phases. 13 Those results also showed an increase in the ovu-

Emotion recognition
Recently, some studies have evaluated neural modulation in relation to recognition of facial emotion according to the phase of the menstrual cycle and sex hormones. For example, the mediating role of the menstrual cycle stage was evaluated with regard to recognition of emotional expressions among 50 women in the menstrual, preovulatory, ovulatory and luteal phases. 26 The authors of that study found that there was greater accuracy regarding fearful faces during the menstrual phase and suggested a hypothesis regarding estrogen levels in the coding of facial stimuli. The fact that fear recognition is affected differently according to the phase of the cycle suggested that there was neural processing distinct from fear.
A comparative study on healthy women and women with premenstrual dysphoric disorder (PMDD) also investigated emotional recognition in order to determine emotional processing errors in the follicular and luteal phases of the menstrual cycle. 17 In that study, there were no differences in emotional recognition among healthy women but it was noted that women with PMDD showed better recognition of negative faces during the luteal phase. According to the authors of that study, this negative bias might contribute to generation of negative mood states during the luteal phase and might suggest the presence of dysfunction in brain regions whose activity mediates recognition of emotion in facial expression. Those findings corroborated the results from electrophysiological investigations. 13,41 The accuracy of facial emotion recognition was evaluated among 40 healthy volunteers. 32 The results showed that women in the follicular phase had greater accuracy in recognizing emotions of sadness and fear than did women in the luteal phase. Similar results were found using fMRI techniques. fMRI measurements with T1 relaxation were made among 20 women at different stages of the cycle and higher levels of blood oxygenation (BOLD) were found in the right spindle area among women taking oral contraceptives than among women who did not use contraceptives during the follicular phase. 25 In addition, the accuracy of emotional recognition was investigated among 22 women (11 in the follicular phase and 11 in the luteal phase) using fMRI T2 relaxation measurements in relation to amygdala activation, and greater neural activation was found for negative emotions such as fear and disgust in the follicular phase. 35 Similar results were found through a neurobehavioral analysis among 35 women during the follicular and luteal phases. 35 Thus, these authors suggested that there was a significant negative correlation between progesterone levels and performance in emotion recognition, such that there was greater accuracy of socioemotional influence with lower levels of progesterone.

Visual attractiveness
Facial features can serve as a clue to social judgments of multiple human features. In one study, female facial stimuli were used in the ovulatory and luteal phases to ascertain the symmetry of facial attractiveness. It was suggested that the greatest facial attractiveness of women was found in the most fertile phase of the menstrual cycle. 39 In another study, an emotional attractiveness task regarding stimuli consisting of female faces, conducted during the follicular and luteal phases, showed that the attractiveness perceived in the women's faces varied over the course of the menstrual cycle and was greater in the ovulatory phase than in the luteal phase. 47 Facial attractiveness was evaluated among 60 women who were presented with stimuli consisting of female faces. 38 They did not find particular attractiveness in faces in the late follicular or luteal phase, but found higher attractiveness during the ovulatory phase. It is important to note that women with high levels of estradiol are more likely to choose the faces of women who are also in the ovulatory phase. In another study, the preferences of men and women for female faces during the ovulatory and luteal periods were investigated and it was shown that there was greater preference for faces in the ovulatory phase. According to the authors of that study, the preference for facial attractiveness in the ovulatory phase correlated with the hormonal fertility state. 40 Stimuli of male and female faces were used in a study aimed at measuring the facial attractiveness of women in the follicular and luteal phases. It was found that there was a higher attractiveness rate in the follicular phase of the menstrual cycle than in the menstrual and luteal phases. 30 Twenty women in the follicular phase and luteal phase were asked to rate the attractiveness of male faces of Caucasian and Japanese ethnicity. 30 There was greater preference for the faces of Caucasian men during the follicular phase than during the luteal phase. Likewise, the facial attractiveness of 31 women in the follicular and luteal phases was measured through the symmetry of male faces. It was shown that women preferred more symmetrical faces during their peak fertility. 36 In line with previous studies, 43 women in the menstrual, follicular, mid-cycle and luteal phases were evaluated and greater attractiveness towards male faces was found during the late follicular phase. 27 In another study, women were more attracted to faces with a greater trait of masculinity when they were in the late follicular phase, due to the gradual increase in estradiol. 15 The converging evidence suggests that the menstrual cycle, it was shown that TRV and TRA were more prolonged in the luteal phase than in the follicular phase. 21 These results could be attributed to the fluctuating levels of female sex hormones. In another study, the TRV and TRA of 50 women in the menstrual, ovulatory and luteal phases were measured using presentation of stimuli such as flashes of light and sound. 33 It was concluded that TRV and TRA were longer during the menstrual phase and shorter during the luteal phase of the menstrual cycle. The data were contradictory, but the authors of that study pointed out that the prolonged reaction times during the menstrual phase were attributable to the decrease in driven reaction times, due to increases in the levels of fluids and electrolytes, and that faster reaction times were due to progesterone, which negates the ability of estrogen to cause a delay in driven reaction time.

Visuospatial cognitive function
Working memory is a faculty of the mind that enables multiple tasks to be performed simultaneously, in parallel. It involves controlling the distribution of concentration, with individual differences in attention span. A study examined visuospatial cognitive functions in the follicular, luteal and menstrual phases of 12 healthy women, with regard to visual working memory. 29 The authors showed that the greatest mental workload and stress was perceived in the luteal phase.

CONCLUSIONS
The results from the studies presented were inconclusive, but these inconsistencies may have been related to lack of experimental control and failure to assess the perimetry of the phases of the menstrual cycle. These findings continue those of Guttridge (1994), although it is difficult to draw specific conclusions about the performance of visual functions during the phases of the normal menstrual cycle. Therefore, further research is needed in order to clarify the sensory and perceptual mechanisms involved in the menstrual cycle, especially with regard to facial recognition and attractiveness processes.
It is noteworthy that while the study by Guttridge (1994) was aimed more towards optical measurements, our study sought to understand neurological and behavioral perceptual processes.