PMS women with sleep problems are commonly characterized by sleep-wake rhythm shifts, subjective sleep disturbance, and sleep electroencephalogram trait variations. In a clinical case report, patients with PMS show a delayed sleep rhythm phase in the luteal phase of the menstrual cycle, whereas the phase is advanced in the follicular phase (71). In addition, PMS-related sleep performance can be unrefreshing and insufficient, which causes daytime sleepiness (72, 73). Although the exact mechanism behind PMS and the sleep-wake cycle is still unknown, elevated progesterone levels during the luteal phase might play a key role since they can induce an increase in body temperature, which will ultimately lead to more fragmented sleep (74, 75).

Sleep disturbances during the luteal phase are often a complaint in women with PMS, including poor sleep, insomnia symptoms, and daytime sleepiness (76, 77) ( Table 1 ). In a study of 127 medical students with PMS, 96 of them (75.6%) struggled with decreased sleep quality (78). A disrupted circadian rhythm and decreased melatonin can be found in sleep disorders (59). However, diseases such as infertility, polycystic ovary syndrome and dysfunction of the hypothalamic-pituitary-ovarian axis, and even an irregular menstrual cycle can also result in sleep disturbances and self-reported poor sleep since dysmenorrhea and accompanying mood changes (4). In a recent survey investigating the association between menstruating women and sleep quality, PMS patients shows sleep duration declines, and menstruating problems are associated with a higher incidence of insomnia and daytime sleepiness (79). As a result, sleep dysfunction is not limited to PMS, and the assessment of healthy people during the menstrual phase and criteria for those defined as PMS should be considered carefully.

Despite tremendous reports of poorer subjective sleep qualities in the late-luteal phase, women with PMS suggest few objective sleep quality alterations by polysomnographic and quantitative electroencephalogram (EEG) measures. In most cases, objective parameter changes are only found between the follicular and luteal phases. When compared to the follicular phase, shorter rapid eye movement (REM) latency and REM episodes are found during the luteal phase in both PMS patients and controls (80, 81). Interestingly, PMDD patients with depressed mood have EEG patterns similar to those of healthy women but distinguish from the EEG architecture in major depressive disorder (82, 83). Taken together, we mention 2 points here and call on more studies to get in. On the one hand, it is suggested that the absence of altered actual sleep may be a peculiarity in PMS patients with poorer perceived sleep quality. On the other hand, mood changes may not be the only contributor to PMS-induced sleep disturbance; other factors, such as circadian rhythm and melatonin dysfunction can also play a critical role.

From the viewpoint of circadian rhythm, PMS-induced sleep disturbance is a reflection of sleep-wake cycle disorder, which can also impact other biological rhythms. In return, the effects of the circadian system on sleep are well recognized, and internal clock impairment is a leading cause of circadian rhythm sleep disorders (CRSDs) (84). Nevertheless, research on the bidirectional relationship between circadian rhythm and sleep in PMS is not sufficient. Shinohara and his colleagues reported body temperature and sleep rhythm changes in a woman with PMS, indicating circadian rhythm and sleep alterations underlying PMS (71). In a laboratory-based study, sleep durations showed consistency with circadian variation (85). A recent cross-sectional study of university students, 78% of whom report social jet lag (different sleep patterns between weekends and school days), shows that circadian system misalignment due to sleep disturbance is significantly associated with menstrual symptoms (86). A reduction in REM sleep sensitive to menstrual phase changes was also found. These findings, although not direct, imply the attribution of circadian rhythm disorder to PMS-induced sleep disturbance.

In PMS, melatonin levels are decreased at both the follicular and luteal phases (24, 87). In the normal menstrual cycle, no alternation is observed in slow wave sleep (SWS). Melatonin secretion appears to be stable during the two menstrual phases. However, in women with PMDD, increased SWS, prolonged objective sleep onset latency, and reduced stage 2 sleep are functionally associated with reduced melatonin levels, and exogenous melatonin can reverse these changes and improve sleep quality (25, 88). These studies provide direct evidence of a potential interaction between melatonin and PMS-induced sleep problems. Herein, although accumulative studies have shown alterations in stage 2 sleep, SWS and REM sleep, blunted melatonin and circadian temperature rhythm in PMS, the exact mechanism of disrupted circadian clock and melatonin secretion in PMS-related sleep problems has not been elucidated.

Recently, melatonin has been suggested to play a role in sleep circadian regulation and sleep disorder treatment. In zebrafish, melatonin is required for sleep regulation, and sleep duration is significantly reduced by blocking melatonin synthesis (36). Regarding human beings, clinical meta-analyses have shown the therapeutic effects of melatonin for patients with first-degree and secondary sleep disorders (89, 90). Specifically, in CRSDs, melatonin can accelerate entrainment in jet lag, advance phases of circadian rhythms in delayed sleep phase syndrome and increase daytime sleep in shift-work sleep disorder (37). Some melatonin receptor agonists, such as tasimelteon, have already been approved for the treatment of CRSDs (91, 92). These findings provide solid evidence of melatonin’s role in regulating sleep disorders, especially those related to circadian system interruption ( Table 2 ).

Melatonin’s effects on sleep disturbance can be found in PMS and other diseases that need differentiation like primary sleep disorder, polycystic ovary syndrome, and primary insomnia (93, 94). For example, an exogenous supply of melatonin can improve sleep quality among women with PMDD (25). Patients with polycystic ovary syndrome, a female reproductive disease with sleep problems, show significantly increased sleep quality after 12 weeks of melatonin supplementation (95). In insomnia, one of the major complaints in women with PMS, melatonin administration can be adopted as a safe adjuvant therapy (96). The therapeutic effects of melatonin may involve the modulation of the internal clock system, as the clock 3111T/C gene polymorphism is associated with altered melatonin levels in women with insomnia (97).

Melatonin’s effects on sleep are diverse depending on the distinct functions of melatonin receptors including MT1 and MT2 (98, 99) ( Table 2 ). In MT1 receptor knockout mice, reduced REM sleep and significantly increased non-REM (NREM) sleep are found whereas MT2 receptor knockout mice present a decrease in NREM sleep (38, 39, 100). In addition, distinct localizations of MT1 (mainly on REM-related regions such as the lateral hypothalamus) and MT2 (mainly on NREM-related areas such as the reticular thalamus) receptors also suggest that melatonin plays receptor-specific roles in sleep regulation (101). Moreover, the MT1 receptor can elevate the amplitude of the internal clock to induce the switch from wake to sleep, and the MT2 receptor may be helpful in synchronizing the sleep cycle to the circadian clock (102). However, the distribution of melatonin receptors has not been elucidated in PMS/PMDD. This evidence provides a molecular basis for melatonin, sleep and circadian clock interactions. Given the circadian rhythm changes in PMS, targeting melatonin and its receptors may relieve PMS-induced sleep problems. Well-conducted and large clinical trials are needed. Relevant mechanisms should also be further explored.

PMS and PMDD patients are characterized by affective disorders such as anxiety and depression. In particular, psychological discomforts such as depression can be the major complaint in women with PMDD (103). As a result, alleviating emotional symptoms is crucial in PMS. Currently, anti-depressants, selective serotonin reuptake inhibitors (SSRIs), have been regarded as the preferred choice in PMS/PMDD treatment (5). However, the dose-dependency and tolerant peculiarities of traditional antidepressants limit their extensive clinical application (104). In recent years, melatonin has increasingly been suggested to regulate emotional symptoms in PMS/PMDD. In PMDD patients, altered circadian rhythms of melatonin are associated with depressed mood (24, 25, 87) ( Table 2 ). In a double-blind study, melatonin administration attenuated premenstrual-like symptoms (41). The authors also found that women treated with melatonin exhibited less depression, anxiety, anger, and fatigue than the placebo control group. Some melatonin receptor agonists, such as agomelatine and ramelteon, have been used as novel anti-depressants and have emerged as promising prospects in depression treatment (105, 106). Together, these studies demonstrate melatonin’s regulation of negative mood, providing a basis for melatonin-mediated emotional changes in PMS.

As the synchronizer of melatonin rhythm, the circadian clock is also involved in emotion regulation. To date, major depressive disorder, bipolar disorder and other affective disorders have been associated with a dysfunctional internal clock system (107–109), which might be attributed to clock gene changes. In a mouse model of depression, Period1(Per1) levels are positively correlated with the severity of depression (110). Per2, brain and muscle ARNT-like 1(Bmal1) and nuclear receptor subfamily 1 group D member 1(Nr1d1) changes also induce depression-like behaviors via inflammation-related processes (57, 111, 112). However, whether the circadian gene Nr1d1 is associated with depressive changes, especially in females, is not fully understood (58, 111, 113). In cryptochrome circadian regulator 1/2(Cry1/2) knockout mice, elevated anxiety levels are observed compared to those in wild-type mice (114). Given that early wake therapy can improve depressive moods without melatonin alternation and that the suppression effects of light on melatonin are compatible between PMDD and healthy people (21, 87), the circadian system and its coupling pathway dysfunction may be included in the etiology of PMS mood disorders.

Apart from its direct anti-anxiety/depression effects, melatonin can also regulate core circadian genes involved in emotional disorders ( Table 3 ). For instance, melatonin increases Per2, Cry1, and Cry2 expression within the anterior pituitary of rats (115). Phase-delayed Per1, Per2, and Cry1 levels are concomitantly reported after melatonin treatment. Moreover, in a seasonal affective disorder mouse model, melatonin supplementation elevates the rhythm amplitudes of Per1, Per2, and Bmal1 in the SCN (116). In addition, scientists have found that melatonin receptor agonists relieved patients’ depression and anxiety with decreased Per1, Per2, Cry1, Cry2 and Nr1d1 expression (28). Melatonin can upregulate Bmal1 level and promote cellular survival by PI3K/AKT signaling pathways (60). These results suggest that melatonin-induced clock gene alterations may be another potential mechanism in PMS-related affective complaints. However, there is still inconsistency inconsistence with limited recognition of the underlying pathways that govern them.

PMS patients often complain of cognitive-related problems in the luteal phase, such as affective lability and a sense of being controlled or overwhelmed. In major depressive disorder, cognitive deficits are proverbially recognized (117), indicating potential cognitive dysfunction in PMS and PMDD. Keenan etal. (6) employed the Trail Making Test B to assess attention capacity differences between PMS and heathy women, and they found that PMS patients exhibited worse performance in the Trails B task during the luteal phase. A recent study compared working memory discrepancies in PMS participants during the follicular phase, which were detected through the N-back task (7). The results suggest that poorer working memory is correlated with increased PMS severity. In PMDD studies, women display poorer N-back task performance in the luteal phase, which is also correlated with PMDD severity, irritability, and functional impairment (118). Admittedly, current studies on PMS/PMDD cognitive alterations are not consistent (119), and some studies fail to report cognitive impairment, especially in PMS patients, which may be due to the decreased severity compared to PMDD. Further studies on cognitive changes in PMS and PMDD should be conducted.

The potential mechanisms of PMS-induced cognitive damage may be attributed to a negative emotional state and dysregulated ovarian steroid secretion. A substantial body of literature has shown that negative moods such as anxiety and depression can result in cognitive impairment (120). Given the characterized affective dysfunction in PMS/PMDD, anti-depression therapy may attenuate cognitive symptoms. Regarding ovarian hormones, estrogen supplementation enhances memory task performance, while progesterone protects cognition after brain injury (121, 122). In addition, estrogen and progesterone receptors are also distributed in brain cognition-related regions, such as the prefrontal cortex and hippocampus. In summary, current studies have demonstrated the emotion and hormone regulation of cognition, although it is plausible whether mood and steroids contribute to cognitive disorders in PMS/PMDD.

Melatonin plays an important role in cognitive regulation ( Table 2 ). In zebrafish, melatonin treatment mitigates cognitive disorder, which results from altered circadian rhythm (47). In a 5-fluorouracil-induced cognitive deficits model, melatonin administration reversed rat spatial memory dysfunction (46). Similar effects of melatonin were discovered in an isoflurane-induced mouse model, and the improved cognitive function after melatonin treatment may be mediated through circadian clock resynchronization (29, 123). In clinical studies, melatonin alleviates cognitive disturbance resulting from breast cancer chemotherapy (124)). Recent advances have regarded melatonin as an index of cognitive impairment in elderly individuals (125). Moreover, melatonin has been shown to effectively improve cognitive deficits in AD mouse models (126). These studies from animals to humans demonstrate a direct regulation of cognition by melatonin. In particular, melatonin has been suggested to play a role in cognition through the regulation of circadian hormones such as estrogen. For instance, nonylphenol, an estrogen mimic, can induce cognitive impairment in Wistar rats, whereas melatonin treatment attenuates its neurotoxicity and adverse cognitive impact (48). Together, potential mechanisms by which melatonin can regulate emotion through circadian clock-dependent or clock-independent pathways have emerged, which may be a possible approach to melatonin regulation of PMS/PMDD-induced cognitive impairments.

PMS symptoms cyclically occur in the luteal phase and gradually vanish after menstruation, indicating the role of the menstrual cycle in PMS occurrence. In line with this viewpoint, one of the most effective treatments for PMS is ovulation suppression, and the combination of estrogen and progestogen shows a promising effect on attenuating PMS symptoms (9). ESC/E(Z) complex genes and ovarian hormone regulation genes also manifest different expression levels in PMDD patients and controls (8). With the changes in ovarian steroids and the subsequent resting regional cerebral blood flow in PMDD, ESC/E(Z) genes seem to have a greater correlation with brain function and more attention should be given (127). More importantly, extensive studies have demonstrated the effects of estrogen and progestogen on depression, anxiety and other emotional disorders in women (128, 129). As a result, ovulation-related hormones, especially estrogen and progestogen, should be considered key factors in the pathogenesis of PMS.

During the menstrual cycle, estrogen and progestogen showed distinct secretion patterns. After ovulation, estrogen levels have a slight drop and last for 1-2 days. Subsequently, both estrogen and progesterone are increased and peak in the mid-luteal phase and are significantly reduced to their lowest levels before menstruation (130). However, it seemed that the levels of ovarian hormones in PMS patients were not similar to those in normal people. Decreased estrogen and elevated progesterone levels may be characteristic of PMS, albeit with inconsistent findings. Recently, Yen etal. (131) thoroughly explored estrogen and progesterone levels in women with PMDD. They found that PMDD patients have decreased estrogen levels in the luteal phase. Moreover, in women with low estrogen levels, elevated progesterone levels are reported in PMDD patients rather than healthy controls. Other independent studies have also confirmed the role of higher progesterone and lower estrogen in PMDD (132, 133). In addition, ovarian steroid fluctuations indicate their interactions with PMDD symptoms. In premenstrual women, more severe premenstrual symptoms are correlated with a reduction in ovarian estrogen and progesterone levels (134). The estrogen and progesterone changes from low to peak are also associated with PMDD onset (135). Remarkably, estrogen receptor alpha (ERα) has gradually received wide attention in PMDD emotional symptoms. Scientists have discovered associations between ERα single nucleotide polymorphisms, the risk of PMDD and patient psychological traits (136–138). The ESR α-Xbal polymorphism also suggests its modulation of PMDD patient emotion (139). In conclusion, these studies demonstrate ovarian hormone-related changes in PMS and indicate their contribution to PMS symptoms from various perspectives. Notably, the production of reproductive hormones involves extensive sophisticated mechanisms and is under the control of the neuroendocrine system. As a result, ovarian hormone-targeted treatment does not simply correct the abnormalities in estrogen and progesterone levels, and hypothalamus-pituitary-gonadal cyclicity and circadian rhythms should also be considered.

Apart from estrogen and progesterone, PMS also induces oscillations of other endocrine factors whose secretion is dominated by hypothalamic–pituitary axes and regulated by the internal clock, such as cortisol, luteinizing hormone (LH), and follicle-stimulating hormone (FSH). Interestingly, estrogen and progesterone may also potentially contribute to these hormone circadian changes. Estrogen is known for its negative regulation of LH and FSH. Allopregnanolone, a metabolite of progesterone, is elevated in PMDD patients with decreased cortisol levels, implying the control of progesterone in cortisol diurnal secretion (140). Although there is no consensus on cortisol circadian rhythm dysfunction, the cortisol awakening response is significantly attenuated in PMS women, suggesting hypoactivity in the hypothalamus-pituitary axis (141, 142). Other hormone circadian changes are also observed in PMS patients, including elevated mean FSH, advanced FSH rhythms and reduced amplitude of LH pulses in the luteal phase of menstruation (143, 144). In summary, changes in estrogen and progesterone seem unpredictable among PMS/PMDD patients. Studies on upstream hormones of ovarian steroids may help us to understand the mechanisms of PMS/PMDD.

Accumulating studies have demonstrated the role of melatonin in ovarian physiology and hormone regulation. In addition to the pineal gland, melatonin is also produced by granulosa cells, the cumulus oophorus, and oocytes, leading to higher concentrations of melatonin in follicular fluid than in serum, which may correlate with estradiol and progesterone levels (145). To date, melatonin has been shown to regulate estrogen and progesterone ( Table 2 ). Normally, decreased estrogen and increased progesterone levels are commonly reported after melatonin interference (146, 147), which can exert its actions through the regulation of enzymes participating in steroid synthesis. Martínez-Campa and his colleagues found that melatonin inhibits aromatase cytochrome p450 (CYP19) mRNA expression, preventing aromatase-induced estrogen production (32). Under melatonin deprivation (pinealectomy), the expression of CYP17A1, a crucial enzyme in estrogen synthesis, is significantly increased in mice (61). However, melatonin treatment promotes steroidogenic acute regulatory protein (StAR) and CYP11A1 expression in pregnant mice, which catalyzes the synthesis of progesterone (31). Moreover, a recent in vitro study also revealed that StAR levels are upregulated after melatonin treatment, leading to elevated progesterone production (33). On the other hand, melatonin can also serve as an estrogen and progesterone receptor modulator. In rats, melatonin treatment contributes to an increase in progesterone receptor A and a decrease in ERα (49). These effects may be mediated via the MT1 receptor, as it is significantly increased in the ovary and can downregulate ERα expression (148). A complementary study was conducted by Talpur etal. (62) They demonstrated that MT1 receptor knockout mice exhibit higher estradiol and lower progesterone levels, indicating the role of the MT1 receptor in melatonin’s modulation of ovarian steroids.

Another potential mechanism involving melatonin effects on ovarian hormone levels is circadian clock regulation within both the hypothalamus and ovary. In the SCN, a timing signal from the central circadian clock is indispensable for the LH surge (149). Given melatonin’s modulation of the central clock, it is reasonable to assume that melatonin may regulate LH levels through the SCN-dependent pathway. Kisspeptin neurons within the hypothalamus also regulate LH secretion by detecting estrogen levels and internal clock signals. Although gonadotropin-releasing hormone (GnRH) neurons control reproductive hormone production, melatonin appears to synchronize ovarian hormone secretion through kisspeptin neurons (63). By targeting the kisspeptin neural system, melatonin induces hypofunction of the hypothalamus-pituitary-ovarian axis, which results in decreased LH secretion and ultimately estrogen level reduction. In the ovary, the circadian clock also suggests its modulation to steroids. Direct evidence is that Bmal1 knockout mice show blunted progesterone levels (64, 65). Moreover, critical genes in steroid synthesis, such as CYP19 and StAR, are also tightly controlled by the circadian clock (31, 66). In light of melatonin’s regulation of clock genes, clock-targeted regulation of melatonin may play an important part in estrogen and progesterone production.

In PMS, estrogen and progesterone levels show incongruent results, although the tendency of estrogen reduction and progesterone elevation is increasingly recognized. Interestingly, PMS/PMDD patients often have decreased melatonin production, which is thought to result in elevated estrogen and decreased progesterone levels. These contrarian changes indicate that the regulation of steroids involves multiple biological processes. One potential explanation is that the effects of melatonin reduction are not dominant compared with those of primary ovarian hormone changes, which can also explain the indeterminate estrogen and progesterone changes in PMS. On the other hand, although estrogen and progesterone do not show direct modulation of the circadian system, the SCN, kisspeptin neurons and ovarian circadian clock genes are clearly regulated by hormone feedback (149), indicating the participation of a disrupted internal clock and other endocrine hormones. Studies on melatonin, the circadian system and ovarian steroid interactions in PMS require further exploration.