Lua Labs Report — Blue light, melatonin suppression, and the menstrual cycle

Date: 2026-06-16 Researcher: Lua Labs Classification: Chronobiology Line: L4 — Hormonal chronobiology Subtopic: 4.4 — Blue light and melatonin suppression: quantifiable impact on the menstrual cycle

External sources

1. Hu, J., Li, S., Yu, X., & Dai, L. (2026). "From nighttime light exposure to menstrual health: a critical review of evidence, mechanisms, and nursing interventions." Frontiers in Reproductive Health, 8:1738574. https://doi.org/10.3389/frph.2026.1738574
2. Vidafar, P., McGlashan, E. M., Burns, A. C., Anderson, C., Shechter, A., Lockley, S. W., Phillips, A. J. K., & Cain, S. W. (2024). "Greater sensitivity of the circadian system of women to bright light, but not dim-to-moderate light." Journal of Pineal Research, 76(2):e12936. https://doi.org/10.1111/jpi.12936
3. Sanchez-Cano, A., Luesma-Bartolomé, M. J., Solanas, E., & Orduna-Hospital, E. (2025). "Comparative Effects of Red and Blue LED Light on Melatonin Levels During Three-Hour Exposure in Healthy Adults." Life, 15(5):715. https://doi.org/10.3390/life15050715
4. Schöllhorn, I., Stefani, O., Lucas, R. J., Spitschan, M., Slawik, H. C., & Cajochen, C. (2023). "Melanopic irradiance defines the impact of evening display light on sleep latency, melatonin and alertness." Communications Biology, 6:228. https://doi.org/10.1038/s42003-023-04598-4
5. Fazlali, F., Lazar, R., Yahya, F., Stefani, O., Spitschan, M., & Cajochen, C. (2025). "Sex and Seasonal Variations in Melatonin Suppression and Alerting Response to Light." Journal of the Endocrine Society, 9(12):bvaf155. https://doi.org/10.1210/jendso/bvaf155
6. Grant, L. K., Gooley, J. J., St Hilaire, M. A., Joffe, H., Brainard, G. C., Van Reen, E., Rüger, M., Rajaratnam, S. M. W., Lockley, S. W., Czeisler, C. A., & Rahman, S. A. (2024). "A pilot study of light exposure as a countermeasure for menstrual phase-dependent neurobehavioral performance impairment in women." Sleep Health, 10(1S):S34-S40. https://doi.org/10.1016/j.sleh.2023.08.012
7. Gooley, J. J., Chamberlain, K., Smith, K. A., Khalsa, S. B. S., Rajaratnam, S. M. W., Van Reen, E., Zeitzer, J. M., Czeisler, C. A., & Lockley, S. W. (2011). "Exposure to room light before bedtime suppresses melatonin onset and shortens melatonin duration in humans." Journal of Clinical Endocrinology & Metabolism, 96(3):E463-E472. https://doi.org/10.1210/jc.2010-2098
8. Greendale, G. A., Witt-Enderby, P. A., Karlamangla, A. S., et al. (2020). "Melatonin Patterns and Levels During the Human Menstrual Cycle and After Menopause." Journal of the Endocrine Society, 4(11):bvaa115. https://doi.org/10.1210/jendso/bvaa115
9. Brown, T. M., Brainard, G. C., Cajochen, C., Czeisler, C. A., Hanifin, J. P., Lockley, S. W., Lucas, R. J., Münch, M., O'Hagan, J. B., Peirson, S. N., et al. (2022). "Recommendations for daytime, evening, and nighttime indoor light exposure to best support physiology, sleep, and wakefulness in healthy adults." PLOS Biology, 20(3):e3001571. https://doi.org/10.1371/journal.pbio.3001571

Baseline knowledge (what I know before searching)

Nighttime light is not only a "visual stimulus." For the circadian system, it is an environmental hormone that enters through the retina. Intrinsically photosensitive retinal ganglion cells (ipRGCs) express melanopsin, with maximal sensitivity around 480 nm; this is why blue-enriched light and many LED screens have disproportionate biological potency relative to their visual appearance. The retina projects through the retinohypothalamic tract to the SCN; the SCN regulates sympathetic output toward the pineal gland through the PVN, intermediolateral column, and superior cervical ganglion. When there is sufficient darkness, pineal norepinephrine activates AANAT and ASMT to convert serotonin into melatonin. When there is nighttime light, that chain is blocked or delayed.

The important point for Lua Labs: melatonin is not just a "sleep signal." In L4.2 we defined it as the endocrine darkness signal. Its amplitude and phase organize temperature, sleep, pancreatic sensitivity, alertness, nocturnal cortisol, and probably reproductive tone through SCN-kisspeptin-GnRH-LH/FSH pathways and local ovarian actions. If a woman has enough hours in bed but receives melanopic load before sleep, she can preserve "sleep duration" and lose "biological night." That difference explains why many sleep-cycle correlations look noisy: the variable is not only sleeping, but entering sleep with intact melatonin.

Measurement also matters. Photopic lux measures brightness for human cones, not circadian potency. Two lights with 80 lux can have different hormonal effects if one concentrates energy at 464 nm and another at 631 nm. The more useful metric is melanopic EDI/mEDI or, in simplified behavioral research, estimated nocturnal melanopic load based on timing, duration, distance, brightness, screen color, ambient light, and prior darkness context.

Before searching the papers, my model was: L4.3 defined the food-melatonin-phase collision. L4.4 should define the other arm: not eating under melatonin, but light that prevents or delays melatonin. The working hypothesis is that screens/blue light do not "alter the cycle" directly and linearly; they erode the darkness signal that synchronizes the SCN, corpus luteum, temperature, sleep, and metabolic response. In the luteal phase and perimenopause, that erosion should be felt more, because progesterone, allopregnanolone, vasomotor load, and temperature are already stressing the night.

Findings from recent papers

Recent literature confirms that the problem should be quantified by spectrum, intensity, and timing. Sanchez-Cano et al. 2025 exposed 12 adults to red light at 631 nm or blue light at 464 nm for three hours, from 21:00 to midnight, with hourly saliva sampling. After 1 h both conditions lowered melatonin, but at 2 h the difference appeared: blue light maintained melatonin at 7.5 pg/mL, while red light allowed recovery to 26.0 pg/mL (p = 0.019); at 3 h the pattern persisted. Schöllhorn et al. 2023 was more precise: in 72 healthy men, manipulating melanopic irradiance from screens without relying only on luminance/apparent color, higher mEDI increased sleep latency, reduced melatonin concentration, and delayed melatonin onset. In other words, cosmetic "night mode" is not enough if melanopic load remains high.

Vidafar et al. 2024 provides the sex/phase piece. In 56 adults, 29 women free of hormonal contraception and 27 men, exposed to 10, 30, 50, 100, 200, 400, and 2000 lux, women had greater melatonin suppression than men at 400 and 2000 lux, but not at 10-200 lux. The finding corrects a dangerous intuition: sensitivity did not differ by menstrual phase and was not acutely associated with E2/P4/testosterone. L4.4 should not say "the luteal phase suppresses more melatonin under light"; it should say "the luteal phase probably pays a higher price for the same melatonin loss because of its thermal-neuroendocrine context."

Fazlali et al. 2025, with 48 adults in a within-subject experiment, found that women showed greater melatonin suppression under moderate light (+4.69%) but a lower alerting response (-6.00%) vs men; winter increased sensitivity to non-visual light effects, with melatonin suppression +18.05% and alertness +7.60% vs summer. They also reported earlier DLMO in the luteal phase than in the follicular phase. This introduces two relevant modulators: season/photoperiod and biological phase, not only clock time.

Grant et al. 2024 shows that nighttime light sufficient to suppress melatonin has phase-dependent neurobehavioral effects. In 29 premenopausal women, with 6.5 h of nocturnal monochromatic light (420-620 nm, 1.03-14.12 uW/cm2), when melatonin was not suppressed, women in the follicular phase had worse reaction time than luteal women; when melatonin was suppressed, the follicular deficit disappeared. This is relevant but delicate: in night work, suppressing melatonin may improve acute performance; in normal life, repeating it may degrade the nocturnal signal. Light should not be moralized: the same stimulus can be useful as an acute work countermeasure and disruptive as a chronic domestic pattern.

Gooley et al. 2011 remains the quantitative human anchor: room light <200 lux before sleep delayed melatonin onset in 99% of individuals, shortened melatonin duration by ~90 min, and, during usual sleep hours, suppressed melatonin >50% in most trials. Brown et al. 2022 translates this into a practical metric: at least 250 melanopic lux at the eye during daytime, maximum 10 mEDI in the 3 h before sleep, and <1 mEDI during sleep. These thresholds are useful as risk bins for estimating whether a night had an intact, eroded, or collapsed dark signal.

Full molecular/endocrine mechanism

The main pathway is retina-SCN-pineal:

Blue-enriched nighttime light / screen / domestic LED
  ↓
ipRGC melanopsin (OPN4, peak ~480 nm) + cone/rod contribution
  ↓
Retinohypothalamic tract → SCN
  ↓
PVN → intermediolateral column → superior cervical ganglion
  ↓
Pineal norepinephrine ↓
  ↓
Tryptophan → serotonin → AANAT ↓ → N-acetylserotonin → ASMT → melatonin ↓/delayed
  ↓
Low MT1/MT2 nocturnal signaling + smaller core temperature drop + alertness ↑

The reproductive bridge does not require light to "touch" the ovary directly:

Low or delayed nocturnal melatonin
  ↓
Desynchronized SCN outflow + short biological-night signal
  ↓
Less robust kisspeptin/GnRH/LH timing + nocturnal cortisol/alertness ↑
  ↓
Less coherent ovarian CLOCK/BMAL1
  ↓
StAR/CYP11A1/HSD3B/CYP19A1 in a less prepared window
  ↓
P4/E2 can be "normal" in point quantity, but with lower temporal amplitude

In the luteal phase, symptomatic cost appears through collision with progesterone and thermoregulation:

Luteal phase → progesterone ↑ → basal temperature ↑ → ALLO/GABA-A modulation
                                           ↓
Nighttime light → melatonin ↓ + delayed phase + alertness/sympathetic tone ↑
                                           ↓
Lower nocturnal heat dissipation + HRV ↓ + awakenings ↑
                                           ↓
Luteal insomnia, irritability, low AM energy, non-restorative sleep

In perimenopause:

Women 42-52 → intermittent ovulation + variable P4 + vasomotor load
       ↓
Melatonin age/stage ↓ + nighttime light/screen lowers amplitude further
       ↓
Narrow thermoneutral zone + nocturnal sympathetic tone + heat awakenings
       ↓
Dark-signal collapse becomes measurable as high nocturnal melanopic load

The interaction with L4.3 is counterintuitive. If blue light delays melatonin, a late dinner might look "less under melatonin" by biological clock, but the result is not protective: the endocrine night shortens, sleep is delayed, DLMO shifts, and metabolic phase moves. That is why analysis of the food-melatonin collision needs a new input: it is not enough to know last-meal time and bedtime; it is necessary to estimate whether melatonin was intact, delayed, or collapsed by light.

The original connection that emerges today: the public variable "screens before bed" is too superficial. The scientific variable is nocturnal melanopic load relative to hormonal phase. The same 30 min of screen use does not mean the same thing in early follicular phase, late luteal phase, perimenopause with hot flashes, or a woman with high morning sunlight exposure.

Lua Labs Hypotheses

Hypothesis 38: Nocturnal melanopic load as a modifier of nocturnal physiology

Statement: In women with active cycles and early perimenopause, high nocturnal melanopic load in the 3 h before sleep predicts worse morning energy and more luteal/nocturnal symptoms, even when total sleep duration appears adequate.

Proposed mechanism: High nocturnal melanopic load reduces or delays melatonin through the ipRGC-SCN-pineal pathway. This creates a short biological night. In the luteal phase, progesterone raises temperature and allopregnanolone modulates GABA-A; without sufficient melatonin, the system loses heat dissipation and a calming signal. The consequence should not be sought only in "I slept little," but in non-restorative sleep, awakenings, lower nocturnal heart-rate variability, higher nocturnal heart rate, low AM energy, irritability, and heat/activation awakenings.

Confidence level: Medium-high for melatonin/sleep; medium for effects on luteal symptoms; low-medium for prediction of cycle length without hormonal measurements.

How to validate:

• With a formal study: cohort n=100, 2 cycles, actigraphy/light sensor or exposure log, nocturnal 6-sulfatoxymelatonin in subset n=40, ovulation confirmed by LH/PdG. Primary outcome: nocturnal melanopic load vs aMT6s and luteal sleep.

Limitations: Real melanopic lux exposure is hard to measure outside the laboratory. A screen question is an imperfect proxy: brightness, distance, night mode, ambient light, screen size, lenses, and prior light change the effect. The menstrual-cycle link is indirect and requires controlling stress, caffeine, alcohol, pain, apnea, shifts, and contraceptives.

Hypothesis 39: Double nocturnal collision — light delaying melatonin + late eating

Statement: The combination of high nocturnal melanopic load and late last intake predicts worse sleep/energy than either one alone, because light shifts the melatonin signal while food imposes metabolic load on a biologically compressed night.

Proposed mechanism: In the food-melatonin collision, the problem is food under melatonin. L4.4 adds a second pattern: screen/intense light delays melatonin, pushes sleep later, and shortens melatonin duration. If there is also a late snack/dessert/alcohol, the pancreas, liver, and intestine operate under an ambiguous circadian signal: neither a clear endocrine night nor full daytime metabolism. This increases latency, postprandial temperature, nocturnal glucose, autonomic arousal, and awakenings.

Confidence level: Medium — the components are well supported; synergy between nighttime light and late eating on hormonal symptoms is an original hypothesis.

How to validate:

• With a formal study: n=60, 4-week natural crossover with wearable + optional CGM + light sensor; endpoints: nocturnal glucose AUC, HRV, latency, and aMT6s.

Limitations: Behavioral patterns cluster: someone who uses a screen late may also work late, eat late, be stressed, or have children. Within-person modeling is needed to separate trait from night-specific exposure.

Hypothesis 40: Perimenopausal dark-signal fragility

Statement: In women 42-52, nocturnal melanopic load has greater symptomatic impact when it coincides with vasomotor load or unstable nocturnal temperature, and it may precede the perception of "my cycle is becoming irregular" before overt menstrual irregularity.

Proposed mechanism: Early perimenopause combines irregular ovulation, intermittent P4, fluctuating E2, a narrow thermoneutral zone, and sympathetic activation. Nighttime light reduces the signal that normally helps cool and consolidate the night. If there are hot flashes/sweating or high skin temperature, high nocturnal melanopic load can act as an amplifier: it does not cause perimenopause, but it makes nocturnal collapse more visible.

Confidence level: Medium — strong mechanistic plausibility, scarce direct longitudinal evidence in perimenopause.

How to validate:

• With a formal study: n=120 perimenopausal women, 6 months, wearable, light/screen diary, vasomotor symptoms, aMT6s in a subset, and hormonal confirmation in a subgroup.

Limitations: Vasomotor load depends on E2, thermal environment, alcohol, weight, apnea, medication, and stress. Nocturnal melanopic load should not become a single explanation or diagnostic tool.

Candidate formulation (if applicable)

Compounds: No compound, medication, or supplement formulation is proposed. The output is a research formulation: Nocturnal Light Context Map.

Target population: Women with active cycles and variable luteal sleep, women with possible PCOS/insulin resistance and high nocturnal digital exposure, and women in early perimenopause with awakenings/heat. In postmenopause, the reading would be sleep/metabolism/vasomotor, not cycle.

Complementary mechanisms: Light load, melatonin-progesterone synchrony, food-melatonin collision, and sleep/food/phase coherence form a map of endocrine night, not a therapeutic recommendation.

Requires validation: First, within-person correlation with sleep/symptoms for 90 days. Second, a study with light sensor + aMT6s before claiming that the proxy estimates melatonin.

Individual variability

Light response varies by age, sex, season, prior light, retina, and chronotype. Vidafar 2024 suggests greater female sensitivity to bright light but not to dim/moderate light; Fazlali 2025 shows that winter may increase light sensitivity. This means rigid universal thresholds should not be used. A medium-brightness screen in winter, after an indoor day without sunlight, can weigh more than the same screen after a day with high natural light.

Genetically, variants in OPN4, PER3, CLOCK, ARNTL/BMAL1, CRY1/2, and MTNR1B could modulate light sensitivity, circadian phase, melatonin response, and nocturnal metabolism. Women with a strong drop in AM energy after late screens, high latency, and repeatedly delayed sleep midpoint could be "light-sensitive." If they also have insulin resistance, the metabolic cost of a shifted night may be greater.

Hormonal stage redefines the cost. In women with active cycles, high nocturnal melanopic load may appear as PMS/luteal insomnia and a more variable cycle. In PCOS, it may amplify chronometabolic misalignment by sustaining late eating and delayed sleep. In perimenopause, the same pattern may appear as heat awakenings, nocturnal anxiety, and AM fatigue. In postmenopause, the reading would be sleep/metabolism/vasomotor, not cycle.

There is also social variability. Urban LATAM normalizes late dinner, phone in bed, illuminated bedrooms, and long workdays. The scientific objective is not to blame habits; it is to detect nights when the darkness signal probably did not exist. That distinction avoids turning chronobiology into moralism.