Abstract
The Moon’s motion around the Earth is complex, governed by interactions among several types of cycles that arise from different aspects of its orbit. The 29.53-day synodic cycle occurs as the Moon passes through new-moon and full-moon alignments (syzygies) with the Earth and the Sun. The 27.32-day tropical cycle occurs as the Moon moves back and forth between its northernmost and southernmost positions (standstills) relative to the plane of the Earth’s equator. The 27.55-day anomalistic cycle occurs as the Moon moves back and forth from its nearest and farthest distances from the Earth (perigees and apogees). The Moon’s effects on luminance and gravity at the Earth’s surface are greatest at times of syzygies, standstills, and perigees. Here, based on periodogram analyses of 14.5 patient-year records of a circular, rapid cycling type of bipolar disorder, we show that onsets of mania in a given individual can recur in association with 2 or more lunar cycles simultaneously, and with conjunctions of syzygies of the synodic cycle with standstills of the tropical cycle or with perigees of the anomalistic cycle. The results are consistent with the fact that the Moon’s effects at the Earth’s surface are the result of interactions among all of its constituent cycles, and they highlight the potential importance of long-term longitudinal designs in studies of lunar influence.
The Moon’s motion around the Earth is complex, governed by interactions among several types of cycles that arise from different aspects of its orbit. Chief among those cycles are the 29.53-day synodic cycle, the 27.32-day tropical cycle, and the 27.55-day anomalistic cycle (Figure 1).
Three major constituent cycles in the Moon’s orbit around the Earth. (a) The Earth orbits around the Sun every 365.25 days. As it does so, the Moon orbits the Earth every 27.32 days. (b) Because the Earth has moved in its orbit around the Sun during that period, the interval between successive alignments (syzygies) of the full moon with the Earth and the Sun is 29.53 days, which is the period of the Moon’s cycle in luminance. In contrast, the combined effect of the Sun and the Moon on gravity is greatest every 14.765 days, when the three bodies are aligned at the new moon and at the full moon and cause the highest (spring) tides. (c) Because the plane of the Moon’s orbit around the earth is tilted relative to the plane of the Earth’s equator, it moves back and forth between northern and southern extremes (maximum and minimum standstills) as it circles the earth every 27.32 days. The Moon’s effects on gravity are greatest at those extremes. (d) The elliptical shape of the Moon’s orbit around the Earth causes it to move back and forth between its nearest (perigee) and farthest (apogee) distances from the Earth every 27.55 days. When the Moon is closest to the Earth, its effects on luminance and gravity are greatest. At those times, “supermoons” (perigean syzygies) give rise to the highest (flood) tides. Figure adapted from Figure 1 in Helfrich-Förster et al. (2021).
The 29.53-day synodic lunar cycle is the familiar ~monthly cycle during which the Moon passes through its 2 points of alignment with the Earth and the Sun. During that cycle, we see a new moon when the Moon passes through the point where it lies between the Earth and the sun, and a full moon when it passes through the point where the Earth lies between the Moon and the Sun. Those two events are referred to as syzygies. They are significant not only because they are conspicuous but also because the effects of the Moon on luminance and gravity at the earth’s surface are greatest at their times of occurrence. Thus, each time the Moon passes through full moon and new moon syzygies, it causes gravity at the Earth’s surface to reach its lowest levels, and ocean tides to reach their highest levels (spring tides). The ~half-month interval between successive syzygies and spring tides is referred to as the 14.77-day semi-synodic lunar cycle. The semi-synodic lunar cycle appears to be a basic unit in the temporal organization of some circular, rapid bipolar mood cycles (Wehr, 2018a, 2018b; Avery and Wehr, 2018).
The plane of the Moon’s orbit around the Earth is tilted relative to the plane of the Earth’s equator. Consequently, as the Moon orbits the Earth, it moves back and forth each ~month between the southernmost and northernmost positions relative to the equator, giving rise to the 27.32-day tropical lunar cycle, a cycle partly analogous to the annual cycle in which the noonday Sun appears to move back and forth between the Tropic of Cancer and the Tropic of Capricorn. Thus, the Moon has its own tropics. Because the tropical lunar cycle is a sinusoid, the Moon seems to stand still when it reaches its northernmost and southernmost positions in the sky, which are referred to as maximum and minimum standstills. Standstills are significant because the Moon’s effects on luminance and gravity are highest at their times of occurrence. As the Moon passes through the 2 standstills, it causes gravity to reach its lowest levels and the tide to reach its highest levels during one of the 2 daily (semi-diurnal) tides. Each time the Moon completes a transit through a tropical cycle, its maximal gravity-lowering and tide-raising effect shifts from one of the 2 daily tides to the other. The half-lunar-month interval between successive standstills and highest tides is referred to as the 13.66-day semi-tropical cycle. The semi-tropical lunar cycle appears to be another basic unit in the temporal organization of some circular, rapid bipolar mood cycles (Wehr, 2018a).
The Moon follows an elliptical orbit around the Earth. Consequently, each time the Moon orbits the Earth, it moves back and forth between its nearest and farthest distances from the Earth (perigees and apogees), giving rise to the 27.55-day anomalistic lunar cycle. Perigees are significant because the Moon’s effects on luminance and gravity are highest at their times of occurrence.
Over long intervals, pairs of constituent lunar cycles generate beat cycles as they go in and out of phase with each other. When in phase, they reinforce each other’s effects on luminance and gravity; in anti-phase, they antagonize each other’s effects. Thus, ~14.76-day semi-synodic cycles and ~13.66-day semi-tropical cycles have their greatest effects when they come into phase with one another twice a year near winter and summer solstices (Figure 2a). Semi-synodic and semi-anomalistic cycles have their greatest effects when they come into phase every ~206 days and cause “supermoons” (perigean syzygies) and flood tides (Figure 2b and 2c). If humans respond to lunar cycles, then one would expect their effects on humans to be greatest at times of conjunctions of syzygies of semi-synodic cycles with standstills of semi-tropical cycles and/or perigees of anomalistic cycles (Figure 2a-2c).
Conjunctions of syzygies of the semi-synodic lunar cycle with standstills of the semi-tropical cycle (a) recur every half lunar year in short-term sequences, but every half solar year in long-term averages. Plots a-c show times of occurrence of syzygies of the 14.77-day semi-synodic lunar cycle (a and b), standstills of the 13.77-day semi-tropical lunar cycle (a), and perigees of the 27.55-day anomalistic lunar cycle (c) in actogram formats based on 177.18-day segments of data (a) or 206.70-day segments of data (b and c) that are plotted sequentially beneath one another. (a) One-half lunar year is 177.18 days. The beat period of 14.77-day semi-synodic lunar cycles and 13.77-day semi-anomalistic lunar cycles is 206.70 days. Red rectangles (a) delineate times of occurrence of conjunctions of syzygies of the semi-synodic lunar cycle with standstills of the semi-tropical lunar cycle. Vertical alignments of successive syzygy/standstill conjunctions indicate that they recurred every half lunar year. Repeatedly, after 3 or 4 half-lunar-year intervals, the vertical alignment of conjunctions moved rightward, that is, 1 semi-synodic/semi-tropical cycle later. As a result, over the long term, the average interval between conjunctions is the ~183-day period of one-half of a solar year, whose recurrences are delineated by a dashed diagonal line that shows times of summer and winter solstices. On average, conjunctions of syzygies of the semi-synodic lunar cycle with perigees of the semi-anomalistic lunar cycle recur every 206 ± 15 days near times when the 2 lunar cycles come into phase with each other (b and c). At times of conjunctions, syzygies alternate back and forth between conjunctions with apogees and conjunctions with perigees. The impact on luminance and gravity at the earth’s surface of conjunctions is greatest when syzygies are conjoined with perigees (supermoons) and least when they are conjoined with apogees. Consequently, the times of impactful conjunctions between syzygies and perigees will shift back and forth relative to ~206-day beats between synodic and anomalistic cycles.
In previous longitudinal studies, we found that onsets of mania in patients with a circular, rapid cycling course of bipolar disorder recurred in association with new moon and/or full moon syzygies of the synodic lunar cycle or maximum and/or minimum standstills of the tropical lunar cycle (Wehr, 2018a; Avery and Wehr, 2018; see also Burgess et al., 2021, 2022). In subsequent longitudinal studies, other investigators reported similar findings in physiological variables in healthy individuals. One group reported that menses onsets recurred in association with new moon and/or full moon syzygies of the synodic lunar cycle, maximum or minimum standstills of the tropical lunar cycle, or perigees and/or apogees of the anomalistic lunar cycle (Helfrich-Förster et al., 2021, 2025). Another group reported that delays in the timing of sleep onset and decreases in the duration of sleep were associated with both new moon and full moon syzygies of the synodic lunar cycle (Casiraghi et al., 2021; Rodriguez Ferrante et al., 2025).
Such complexity is not surprising if one considers the complexity of the Moon’s effects on ocean tides. Over time, the effects of interactions of the Moon’s constituent gravimetric cycles cause the amplitude and timing of the tides to vary to a great extent. At the same time, those same interactions, when viewed from a large-scale perspective, give rise to recurring regularities in the tides based on repetitions of their short-term variations. If humans respond to lunar cycles, might the same be true of their response? When viewed from a large-scale perspective, would humans’ responses to the Moon mirror the ordered complexity of the ocean’s responses?
To address that question, we retrospectively analyzed 4 patients’ multi-year, daily records of mood changes that occurred in a circular, rapid cycling form of bipolar disorder. In previous longitudinal studies, we presented evidence that onsets of mania in that type of patient recurred in association with synodic, tropical, and/or anomalistic lunar cycles (Wehr, 2018a, 2018b; Avery and Wehr, 2018). In the present study, we asked whether long-term variations in the timing of onsets of mania would be associated with interactions among the 3 types of lunar cycles.
In our analyses, we used periodograms and eigenvalue pencil tests to determine whether more than one type of lunar cycle was present in the timeseries of each patient’s mood ratings, and whether periods that matched the timing of conjunctions of syzygies of the synodic cycle with standstills of the tropical cycle or with perigees of the anomalistic cycle were present. We then used actogram plots to delineate courses of onsets of mania and to elucidate their phase relationships to courses of syzygies, standstills, and perigees.
Materials and Methods
Patient Selection
We required patients to meet criteria for rapid cycling bipolar disorder that were in use during the eras when they were diagnosed. In addition, we required their illness to follow a circular course in which multi-week episodes of depression regularly alternated with multi-week episodes of mania, consistent with Falret’s 1854 description of la folie circulaire (Falret, 1854). The latter requirement was based on an expectation that lunar cycles would most likely be found in physiological and behavioral phenomena that are known to recur in a cyclic manner, like the menstrual cycle. It is important to note that cases diagnosed as rapid cycling according to the current Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, (DSM-5) criteria would not necessarily meet that requirement (American Psychiatric Association, 2022).
For our analyses, we selected 2 women (patients 2 and 4, ages 69 and 49) and 2 men (patients 1 and 3, ages 58 and 41) for whom multi-year records of their mood cycles were available. Taken together, the records consisted of more than 14.5 patient-years of observations. During the periods of observation, the patients’ symptoms of mania were moderate and did not require hospitalization. The frequency of their manic and depressive episodes ranged from 12 to 28 per year, a number considerably greater than the 4-episode minimum required for diagnosis of rapid cycling by the Diagnostic and Statistical Manual of the American Psychiatric Association, Version 5 (American Psychiatric Association, 2022).
Most of the time, all 4 patients were treated with combinations of antidepressants and mood stabilizers. The onset of mania and rapid cycling in patient 1 occurred during a course of treatment with a fluoroquinolone antibiotic and a travel-related 6-hour phase-advance of his sleep-wake cycle. During the 2-year period when he recorded mood changes, he was treated with a mood stabilizer, lithium carbonate, and an antidepressant, sertraline. Subsequently, after sertraline was discontinued, his mood became more stable. Patient 2 was treated with a mood stabilizer, lithium carbonate, and a tricyclic antidepressant, amitriptyline, during the period when she experienced rapid cycling of mood. When those medications were discontinued, rapid cycling ceased. The onset of mania and rapid cycling in patient 3 occurred after he began treatment with the antidepressant fluoxetine. During most of the period of observation of his mood cycles, he was treated with mood stabilizers lithium carbonate and carbamazepine, the antidepressant/mood stabilizer lamotrigine, and the antidepressant/neuroleptic quetiapine. He received ketamine infusions from late winter to early summer every year. Patient 4 was treated with the tricyclic antidepressant amitriptyline and the mood stabilizer lithium carbonate during the period when she experienced rapid cycling of mood. When amitriptyline was discontinued while lithium carbonate was maintained, rapid cycling ceased.
Patients 2 and 4 participated in research conducted by investigators in the Intramural Research Program of the National Institute of Mental Health at the National Institues of Health Clinical Center in Bethesda, Maryland. The authors saw patients 1 and 3 in consultation, while independent psychiatrists managed their treatment. Patients 1, 2, and 4 correspond to patients 8, 11, and 17 in Wehr (2018a). Patient 3 was not reported previously.
All patients gave written consent for publication of the timeseries of ratings of their mood. Patients 2 and 4 did so as part of their participation in diverse studies that were approved by the Institutional Review Board of the Intramural Research Program of the National Institute of Mental Health, Bethesda, Maryland.
Mood Ratings
Records kept by patients were of 3 types. Patients 1 and 2 recorded dates of switches back and forth between depression and mania. Patient 3 recorded daily totals of symptoms of depression and mania using a self-devised list of symptoms that he considered to be typical of his episodes. Patient 4 recorded daily ratings of mood using a modified version of the Bunney-Hamberg Rating Scale (Bunney and Hamberg, 1963).
Data Processing
For data that patients recorded as dates of switches back and forth between depression and mania, we defined manic and depressive episodes as intervals from the day of a switch to the day preceding the ensuing switch of opposite polarity. We then created Excel files of the timeseries of alternating manic and depressive episodes by entering “1s” for each day of a manic episode and “0s” for each day of a depressive episode.
For data that patients recorded as daily ratings of mood, we created Excel files of the timeseries of alternating manic and depressive episodes by entering values of each day’s mania rating. In those cases, ratings of “0” corresponded to depression. For patient 3, we set a threshold of 3 symptoms for the diagnosis of mania by subtracting 2 from the total number of symptoms that he recorded each day. We then calculated a 5-point moving average of the time series.
From the website, astropixels.com, courtesy of Fred Espenak, we obtained data about times to the nearest minute of full moon and new moon syzygies of the synodic cycle, maximum and minimum standstills of the tropical cycle, and perigees of the anomalistic cycle that occurred during the periods when patients recorded their mood states. We then created Excel files of those variables by entering each day’s data, using “1s” to indicate the occurrence of syzygies, standstills, or perigees, and “0s” to indicate their absence.
Statistical Analysis
To detect periodicities in patients’ mood ratings and in contemporaneous lunar ephemera, we used the ActogramJ plugin for ImageJ to calculate Lomb-Scargle periodograms, with the threshold for statistical significance set at p < .00001 (Schmid et al., 2011; Ruf, 1999; VanderPlas, 2018). We focused on 2 contiguous period-domains: 10.00-100.00 days, an interval that would include periods of the patients mood cycles, as well as periods of synodic, tropical, and anomalistic lunar cycles, and 100-600 days, an interval that would include beat periods of the lunar cycles, lunar years, and solar years. In 2 instances where Lomb-Scargle periodograms failed to detect known periodicities of syzygies, standstills, and perigees, we used χ2 periodograms (Sokolove and Bushell, 1978) instead, as indicated in the figures. In addition, we used Claude Sonnet (Anthropic, claude.ai) to perform eigenvalue pencil analyses to detect periodic elements in the timeseries of patients’ mood ratings.
To facilitate visual inspection of the courses of the patients’ mood cycles and their relationships to courses of syzygies, standstills, and perigees of the synodic, tropical, and anomalistic lunar cycles, we used ActogramJ to create actogram-type plots of mood cycles and contemporaneous lunar ephemera (Schmid et al., 2011).
Results
Patient 1
Periodogram analysis detected 27.39-, 29.01-, 41.83-, and 44.38-day periods in mood ratings that were nearly identical, respectively, to 27.30-, 29.52-, 40.96-, and 44.27-day periods of recurrence of every tropical lunar cycle, every synodic lunar cycle, every third semi-tropical lunar cycle, and every third semi-synodic lunar cycle (Figure 3). If the mood cycles were entrained to the lunar cycles, then the findings would be consistent with 1:1 modes of coupling of mood cycles to the lunar cycles in the first two instances, and 1:3 modes in the second two instances.
Eighteen of 25 periods detected by Lomb-Scargle periodograms in the times series of patients’ mood assessments were nearly identical to integer multiples of periods of lunar cycles (see also Figure 4). Black numerals over peaks in periodograms indicate periods detected in patient’s mood ratings. Red numerals indicate multiples of periods of lunar cycles associated with those peaks.
The periodogram analysis also detected a 351.67-day period in mood ratings that was close to the transient (see Figure 2a) 354.01-day lunar year interval between conjunctions of syzygies of the synodic cycle with standstills of the tropical cycle (Figure 4).
Eighteen of 25 periods detected by Lomb-Scargle periodograms in the times series of patients’ mood assessments were nearly identical to integer multiples of periods of lunar cycles (see also Figure 3). Black numerals over peaks in periodograms indicate periods detected in patient’s mood ratings. Red numerals indicate multiples of periods of lunar cycles associated with those peaks. ~354 days = 12 synodic lunar months (1 lunar year) and 13 tropical lunar months.
In actogram plots of the patient’s mood cycles, vertical alignments of onsets of mania indicated an association of onsets with recurrences of new moon and full moon syzygies of the synodic lunar cycle (Figures 5a and 6a and 6b) and an association with minimum standstills of the tropical lunar cycle (Figure 5d). Associations of onsets of mania with syzygies of semi-synodic lunar cycles and standstills of the semi-tropical cycle occurred near times of syzygy/standstill conjunctions, when effects of syzygies and standstills on luminance and gravity at the earth’s surface were mutually reinforcing (Figure 6a and 6b).
Onsets of mania in patients 1 and 2 recurred in association with syzygies of synodic cycles, standstills of tropical cycles, and interactions between those cycles. Horizontal black and white bars in the actograms indicate manic and depressive episodes, respectively, in patient 1’s (left) and patient 2’s (right) records of mood changes. Vertical and diagonal red lines in a, b, f, and g indicate times of syzygies of the synodic cycle and standstills of the tropical cycle, respectively. Vertical red lines in d and i indicate times of standstills of the tropical cycle. Black and white circles indicate times of new moon and full moon syzygies, respectively. In the periodograms (c, e, h, and j), black numerals denote mood cycle periods, and red numerals denote lunar cycle periods. Periodogram analyses of patient 1’s records detected 27.40-day and 29.01-day periods that were close to the 27.30-day and 29.53-day periods of synodic and tropical lunar cycles, respectively (c), and a 352-day period that was close to the transient 354.61-day interval (see Figure 2a) between successive conjunctions of syzygies of the synodic cycle and standstills of the tropical cycle (e). Periodogram analyses of patient 2’s records detected 54.79-day and 59.26-day periods that were close to 54.64-day and 59.06-day periods of recurrence of every second tropical and every second synodic lunar cycle, respectively (h), and a 183.05-day period that was close to the long-term average 182.52-day interval (see Figure 2a) between successive conjunctions of syzygies of the synodic cycle with standstills of the tropical cycle (j). Note that rapid cycling between mania and depression ceased when treatment with psychotropic medications (tricyclic antidepressant and lithium carbonate) was suspended (g). Start dates: 31 December 2000 (patient 1); 28 January 1980 (patient 2). Figures (a) and (f) adapted from Figures 7a and 3a in Wehr (2018a) and Wehr and Helfrich-Förster (2021), respectively.
Onsets of mania in patients 1 (a) and 2 (c) recurred in association with syzygies of the 14.77-day semi-synodic cycle when syzygies occurred near times of their conjunctions with standstills of the 27.32-day tropical cycle, and their effects on luminance and gravity at the earth’s surface were mutually reinforcing (b and e). Vertical red lines indicate time of syzygies of the synodic cycle and their conjunctions with standstills of the tropical cycle. Vertical red lines in c indicate times of syzygies of the semi-synodic cycle. Diagonal red lines in e indicate times of conjunctions of syzygies with standstills. Diagonal dashed lines in e indicates times of winter solstices. White and black circles indicate times of full moon and new moon syzygies, respectively.
Eigenvalue pencil analyses detected periods in mood ratings that were identical to integer multiples of both semi-tropical and semi-synodic lunar cycles (Table 1).
Eigenvalue pencil analyses detected periods in patients’ mood ratings that were close to lunar periodicities and similar to periods detected in Lomb-Scargle periodograms.
Lunar periods 13.66, 27.32, 40.98, and 54.63 days correspond to periods of recurrence of every single, second, third, and fourth semi-tropical lunar cycle. Lunar periods 14.77, 29.53, 44.3, and 59.06 days correspond to periods of recurrences of every single, second, third, and fourth semi-synodic lunar cycle. Lunar period 354.36 days (the lunar year) correspond to periods of recurrence of every 13th tropical lunar cycle and every 12th synodic lunar cycle.
Patient 2
Periodogram analysis detected 54.81- and 59.29-day periods in mood ratings that were nearly identical, respectively, to 54.64- and 59.06-day periods of recurrence of every second tropical lunar cycle and every second synodic lunar cycle (Figure 3). If the mood cycles were entrained to the lunar cycles, then the findings would be consistent with 1:2 modes of coupling of mood cycles to the lunar cycles.
The periodogram analysis also detected a 183.06-day period in mood ratings that was nearly identical to the long-term average (see Figure 2a) 182.63-day semi-annual interval between successive conjunctions of syzygies of the semi-synodic cycle with standstills of semi-tropical cycles (Figure 4).
In actogram plots of the patient’s mood cycles, vertical alignments of onsets of mania indicated an association of onsets with new moon and full moon syzygies of the synodic lunar cycle (Figures 5f and 6c and 6d) and an association with maximum standstills of the tropical cycle (Figure 5i). Associations of onsets of mania with syzygies of semi-synodic lunar cycles occurred near times of syzygy/standstill conjunctions, when effects of syzygies and standstills on luminance and gravity at the earth’s surface were mutually reinforcing (Figure 6c and 6e).
Eigenvalue pencil analyses detected periods in mood ratings that were nearly identical to integer multiples of both semi-tropical and semi-synodic lunar cycles (Table 1).
Patient 3
Periodogram analysis detected 14.77-, 29.84-, and 40.24-day periods in mood ratings that were nearly, or actually identical, respectively, to 14.77-, 29.53-, and 40.96-day periods of recurrence of every semi-synodic, every synodic, and every third semi-tropical lunar cycle (Figure 3). If the mood cycles were entrained to the lunar cycles, then the findings would be consistent with 1:1 modes of coupling of mood cycles to lunar cycles in the first two instances, and a 1:3 mode of coupling of the mood cycle to the lunar cycles in the third instance.
The periodogram analysis also detected a 356.12-day period in mood ratings that was close to the transient (see Figure 2a) 354.45-day lunar-year interval between conjunctions of syzygies of the synodic cycle with standstills of the tropical cycle (Figure 4).
Z-scores detected statistically significant peaks and troughs in average 29.53-day synodic lunar month, 27.31-day tropical lunar month, and 27.55-day anomalistic lunar month profiles of the patient’s mood ratings (Figure 7b). An association of bouts of mania with new moon and full moon syzygies was most prominent during short days of the year in 354.60-day (lunar year) averages of mania ratings (Figure 7c).
Onsets of mania in patient 3 recurred in association with syzygies of the synodic lunar cycle (b, left, and d), standstills of the tropical cycle (b, center), perigees of the anomalistic cycle (b, right), and with 354.6-day (lunar year) recurrences of conjunctions of syzygies of the synodic cycle with standstills of the tropical cycle (e). Indications as in Figure 5. Periodograms of patient 3’s mania ratings detected 29.55-day and 355.11-day periods that were virtually identical to the 29.53-day and 354.60-day periods of synodic cycles and the transient 354.60-day interval (see Figure 2a) between successive conjunctions of syzygies of the synodic cycle and standstills of the tropical cycle (d and e). Actograms (a) were plotted with successive segments of daily mania ratings that were equal to periods of synodic (left), tropical (middle), and anomalistic (right) lunar cycles. In average profiles of those segments (b), peaks in mania were associated with peaks in recurrences of syzygies of the synodic cycle, standstills of the tropical cycle, and perigees of the anomalistic cycle. In average lunar-year profiles of mania ratings, syzygy-associated peaks in mania ratings occurred predominately during the short days of the year (c). Although the periodogram analysis did not detect 27.32-day tropical cycles or 27.55-day anomalistic cycles in the ratings, statistically significant z-scores (asterisks in b) were associated with troughs and peaks in 27.32-day and 27.55-day average profiles of the ratings, and eigenvalue pencil analyses detected periods close to integer multiples of both semi-synodic and semi-tropical lunar cycles (Table 1). d: χ2 periodograms. Start date: 1 August 2019.
Eigenvalue pencil analyses detected periods in mood ratings that were nearly identical to integer multiples of both semi-tropical and semi-synodic lunar cycles, and to the transient lunar-year period of conjunctions of syzygies of the synodic cycle with standstills of the tropical cycle (Table 1).
Patient 4
Periodogram analysis detected a 55.87-day period in mood ratings that was nearly identical to the 54.67-day period of recurrence of every fourth semi-tropical lunar cycle (Figure 3). If the mood cycle was entrained to the lunar cycle, then the finding would be consistent with a 1:4 mode of coupling of the mood cycle to the lunar cycle.
Although a Lomb-Scargle Periodogram failed to do so (Figure 4), a χ2 periodogram detected a 177-day period in mood ratings that was nearly identical to the transient (see Figure 2a) 177.65-day (semi-lunar year) interval between conjunctions of syzygies of the semi-synodic lunar cycle with standstills of the semi-tropical lunar cycle (Figure 8d).
The majority of onsets of mania in patient 4 (e) recurred in association with standstills of the 27.34-day tropical lunar cycle (vertical red lines in e) and with 177.65-day (half lunar year) recurrences of conjunctions of syzygies of the synodic cycle with standstills of the tropical cycle (d). The periods 55.87 days (c) and 177.00 days (e) detected in the timeseries of daily mood ratings were virtually identical, respectively, to 54.67-day recurrences of every second 27.33-day tropical lunar cycle and transient (see Figure 2a) ~178-day intervals between conjunctions of syzygies with standstills. A 215.50-day period detected in mood ratings (d) was virtually identical to a 214.14-day pattern in the recurrence of syzygies of the synodic cycle (see Figure 9b). A 35.69-day period detected in daily mood ratings (c) is reflected in the vertical alignment of mania onsets in a. A 42.87-day period (c) is reflected in their diagonal alignment of successive mania onsets in b. Simple mathematical relationships among periods detected in the daily mood ratings may indicate that they arise from different perspectives on a common underlying process: 6 (35.69 days) = 5 (42.87 days) = 214.14 days. Note that rapid cycling between mania and depression ceased when treatment with the tricyclic antidepressant medication amitriptyline was suspended (a). d: χ2 periodograms. Start date: 22 October 1974. Figure 4a adapted from the cover of Wehr and Goodwin (1983).
In an actogram plot of the patient’s mood cycles based on the 177.65-day half-lunar-year period (Figure 8e), vertical alignment of 12 of 18 onsets of mania indicated an association of onsets with standstills of the semi-tropical lunar cycle (Figure 8f).
The periodogram analysis also detected a 215.40-day period in the patient’s mood ratings (Figure 4) that was nearly identical to a 214.14-day sequence of 14.5 semi-synodic lunar cycles (Figure 9b). In an actogram plot of the patient’s mood cycles based on that 214.14-day interval, 5 onsets of mania occurred during each repetition of the 214.14-day interval (Figure 9c). With every other repetition of the 214.14-day interval, 4 consecutive onsets of mania recurred in association with syzygies of the 14.77-day semi-synodic lunar cycle; with every intervening repetition, 4 consecutive onsets of mania recurred in anti-phase (Majhi et al., 2020) association with syzygies of the semi-synodic cycle (Figure 9b and 9c). With each repetition of the 214.14-day interval, the fifth onset of mania recurred in association with conjunctions of syzygies of the 14.77-day semi-synodic lunar cycle with perigees of the 27.55-day anomalistic cycle (Figure 9a-9c). At the same time, 3 onsets of mania recurred in association with standstills of the 13.67-day semi-tropical lunar cycle (Figures 8e and 8f; 9d and 9e). In this complex manner, onsets of mania recurred in association with syzygies, standstills, and perigees of the 3 types of lunar cycles in a regularly recurring sequence.
Onsets of mania in patient 4 were associated with complex interactions between ~221-day recurrences of conjunctions of syzygies with perigees (a and b) and ~178-day recurrences of conjunctions of syzygies with standstills (d and e). Vertical line segments indicate times of perigees of the anomalistic cycle in (a) and syzygies of the synodic cycle in b and e. Daily mania ratings are shown in black in c and d. Periodogram analysis of patient 4’s mania ratings (Figure 4) detected a 215.40-day period that was virtually identical to a regular 214.14-day pattern in recurrences of syzygies of the synodic lunar cycle. In that pattern, syzygies shifted 180° with each repetition of the 214.14-day sequence (b). Each 214.14-day period encompassed a sequence of 5 onsets of mania (c and d). With each repetition of the 5-onset sequence, the first 4 onsets alternated every ~214 days between in-phase associations with syzygies of the synodic cycle in 1 repetition and anti-phase associations with syzygies (17) in the next repetition (vertical dashed lines in b and c). The fifth onset in each sequence was associated with conjunctions of syzygies of the synodic cycle with perigees of the anomalistic cycle (diagonal dashed lines in a, b, and c). Diagonal dashed lines in d and e indicate courses of standstills of semi-tropical cycles.
Eigenvalue pencil analyses detected periods in mood ratings that were nearly identical to integer multiples of both semi-tropical and semi-synodic lunar cycles (Table 1).
In summary, Lomb-Scargle periodograms detected periods in mood ratings that were nearly identical to integer multiples (resonant frequencies) of semi-synodic cycles in patients 1-3 and semi-tropical cycles in all 4 patients (Figure 3). Eigenvalue pencil analyses detected periods in all 4 patients’ mood ratings that were nearly identical to integer multiples of both semi-tropical and semi-synodic lunar cycles (Table 1). Lomb-Scargle periodograms also detected ~lunar year or ~half-lunar-year periods that matched transient (see Figure 2a) intervals between conjunctions of syzygies of the semi-synodic cycle with standstills of the semi-tropical cycle in patients 1, 3, and 4, and a half-solar-year period that matched the long-term average interval between conjunctions of syzygies of the semi-synodic cycle with standstills of the semi-tropical cycle in patient 2 (Figure 4). An eigenvalue pencil analysis detected a lunar-year period in patient 3’s mood ratings. In patient 4, a periodogram detected a ~7-month period during which 4 onsets of mania were associated with synodic cycles (alternating between an in-phase relationship with syzygies in one ~7-month sequence and an anti-phase relationship with syzygies in the next ~7-month sequence); the fifth onset of mania in each ~7-month sequence was associated with conjunctions of syzygies of the semi-synodic cycle with perigees of the anomalistic cycle (Figures 8 and 9).
The presence of both vertical and diagonal alignments in the onsets of mania in several of the actograms (Figures 5a, 5b, 5f, and 5g; 9c and 9d) is consistent with the presence of 2 cycles with different periods. Vertical alignment of onsets in the actograms indicates that they recurred with periods similar to the length of the data segments that were used to construct the actograms; diagonal alignment of onsets indicates that their periods of recurrence were shorter, if right leaning, or longer, if left leaning.
In patients 2 and 4, the regular and frequent association of onsets of mania with syzygies of the synodic cycle, standstills of the tropical cycle, and in patient 4, syzygies of the synodic cycle, standstills of the tropical cycle, and perigees of the anomalistic cycle appeared to depend on maintenance treatment with the tricyclic antidepressant medication amitriptyline. When the medication was discontinued, the patients experienced very long periods of hypomania that alternated with shorter periods of depression (Figures 5g and 8a). The periods of hypomania generally were well tolerated.
Discussion
Our findings indicate that onsets of mania in 4 patients recurred in association with more than one type of lunar cycle at a time, and with interactions between pairs (Figure 5) or triplets (Figures 7-9) of different types of lunar cycles. If onsets of mania were entrained to lunar cycles, the complexity in their patterns of recurrence would not be surprising because the Moon’s effect on luminance and gravity at the earth’s surface exhibits a similar complexity, a complexity that arises from interactions among its constituent cycles. Consequently, the multiplicity of cycles in patients’ courses of illness could be taken as further evidence of lunar influence on humans.
Consistent with results of a previous study (Wehr, 2018a), there were differences among individuals, and within individuals over time, in the ratios between periods of recurrence of mania and periods of recurrence of syzygies or standstills. For example, onsets of mania in patient 1 recurred in association with 29.53-day recurrences of syzygies of every synodic cycle (Figures 3 and 5a), while onsets of mania in patient 2 recurred in association with 59.06-day recurrences of syzygies of every second synodic cycle (Figures 3 and 5f). If onsets of mania were entrained to synodic cycles, then differences in the intrinsic periods of each patient’s mood cycles could account for differences in their coupling ratios: mood cycles with longer intrinsic periods would couple to higher multiples of lunar cycles than mood cycles with shorter intrinsic periods would.
Novel findings in the present study were the ~352-, ~183-, ~355-, and ~177-day periods detected in patient 1-4’s respective mood ratings. While the association does not prove causation, those periods are close to transient ~354-day lunar-year and ~177-day half-lunar-year intervals between conjunctions of syzygies of the semi-synodic cycle with standstills of the semi-tropical cycle (see red rectangles in Figure 2a). The 183-day (half solar year) period is identical to the ~183-day beat period of semi-synodic and semi-tropical cycles (see dashed diagonal line in Figure 2a), but also identical to the half solar year. Thus, an alternate or additional seasonal solar influence cannot be excluded in that instance. Note that over ~1.5- to ~2-year periods, intervals between conjunctions of syzygies with standstills can be identical to ~177-day half- or ~354-day full-lunar-year periods, while over longer periods, the average interval between conjunctions is identical to ~183-day half- or ~365-day full-solar-year periods. As shown in Figure 2a, the shift in syzygy/standstill conjunctions’ orientation from a lunar to a solar frame of reference results from occasional delays in the timing of conjunctions from a leading syzygy/standstill pair to a trailing one (see red rectangles in Figure 2a).
Moonlight Versus Moon Gravity
The magnitude of effects of lunar cycles in gravity on an object as small as the human body has been compared to the weight of a mosquito. In light of that fact, lunar cycles in luminance would seem to be a much more likely mediator of the Moon’s effects on humans. However, that supposition fails to take into consideration the impact of modern, urban environments on humans’ exposure to moonlight. In other types of animals, there is now considerable evidence that effects of moonlight on biology and behavior have been disrupted by light pollution in environments where humans live. It seems likely that the same would be true of humans, especially if one considers that humans further isolate themselves from moonlight with architectural barriers. In contrast, the Moon’s effects on gravity in modern human environments are undiminished.
Parallel changes have occurred in humans’ exposure to sunlight. In previous studies, we showed that the human organism can detect and respond to seasonal changes in the timing of sunrise and sunset but fails to do so in modern urban environments (Wehr et al., 1993, 1995). In light of those findings, it is reasonable to infer that humans are even less likely to detect and respond to changes in the timing of moonrise and moonset in those environments.
Even if humans were incapable of detecting lunar cycles in gravity directly, it is not excluded that they could detect such cycles indirectly, via Moon-induced cycles in geophysical variables, such as the geomagnetic field, or some other known or unknown medium. The fact that, until recently, animals were thought to be incapable of detecting the earth’s magnetic field is a cautionary note in the discussion of potential human capabilities.
The results of longitudinal studies of lunar cycles in humans seem more consistent with mediation by lunar gravity than with mediation by moonlight. For example, onsets of menses, sleep, and mania have been reported to recur in association with both full moon and new moon phases of the Moon’s 14.77-day cycle in gravity (Wehr, 2018a; Helfrich-Förster et al., 2021, 2025; Casiraghi et al., 2021; Rodriguez Ferrante et al., 2025). In contrast, the Moon’s luminance cycle is a 29.53-day cycle. A recent finding that gravity at the Earth’s surface predicts sleep timing in 2 populations of humans, and in a population of non-human primates with attenuated access to natural light, provides additional support for gravity as a mediator of lunar influence on humans (Rodriguez Ferrante et al., 2026).
Biological Significance of Lunar Cycles in Humans
Our findings raise a question as to the biological significance of lunar cycles in humans. A common function of such cycles in other animals is to increase reproductive success by synchronizing reproductive behavior and physiology among individual members of a species. The fact that menstrual cycles can become synchronized with synodic lunar cycles, and that lunar cycles in bipolar disorder cycles are associated with dramatic changes in libido seems consistent with such a function in humans. On the other hand, the fact that menstrual cycles and bipolar mood cycles can recur in association with any of the 3 different types of lunar cycles, and with either of the 2 opposite phases of those cycles, seems inconsistent with a function that serves to enhance reproductive success by synchronizing the reproductive behavior of individual members of a species.
To account for that paradox, we hypothesized that humans have redundant biological mechanisms that can detect lunar cycles in both luminance and gravity (Wehr, 2018a). In a natural environment, effects of synodic lunar cycles in luminance would predominate over effects of semi-synodic lunar cycles in gravity, and onsets of menses and mania would maintain consistent phase-relationships with full moon syzygies. Cycles in gravity would play a secondary role by reinforcing synodic cycles in luminance, and by maintaining entrainment to the Moon during periods of overcast skies. In the modern human environment, the controlling influence of synodic lunar cycles in luminance would be lost, and effects of lunar cycles in gravity would predominate over effects of synodic lunar cycles in luminance, resulting in a loss of function.
Biological Mechanism of Lunar Cycles in Humans
While mechanisms that enable humans to detect and respond to lunar cycles remain to be discovered, there are reasons to think that the Moon induces mania in circular, rapid cycling bipolar patients by disrupting sleep: (1) longitudinal observations of healthy individuals show that full moon and new moon syzygies are associated with delayed sleep onset and reduced sleep duration (reviewed in Casiraghi et al., 2021); (2) there is evidence that reduced sleep precedes onsets of mania during the natural course of bipolar disorder (Leibenluft et al., 1996; Ulrichsen et al., 2025); and (3) experimental deprivation of sleep can precipitate mania in depressed, circular, rapid cycling bipolar patients (Wehr et al., 1982). In light of those observations, we hypothesize that rapid cycling of mood in bipolar patients is a pathological response to gravitational lunar cycles in the timing and duration of sleep that occur physiologically in healthy humans (Casiraghi et al., 2021; Rodriguez Ferrante et al., 2026). Our hypothesis is an extension of an earlier, more general hypothesis that sleep reduction is a final common pathway in the genesis of mania (Wehr et al., 1987; Wehr, 1991).
The finding that recurrences of patients’ onsets of mania were associated with integer multiples of 14.77-day semi-synodic cycles seems to indicate that the semi-synodic cycle is a basic unit in the temporal organization of their mood cycles. The 14.77-day semi-synodic cycle arises from interactions between the 12-hour solar component and the 12.42-hour lunar component of the semi-diurnal lunisolar gravimetric cycle. As the 2 components beat against each other, they come into phase every 14.77 days. At those times, the 2 cycles reinforce one another’s effects on gravity, and the tides reach their highest levels. Some organisms that live in the intertidal zone have endogenous circa-12.42-hour biological rhythms that mirror and anticipate the 12.42-hour component of the tidal cycle. Recently, investigators reported that humans, in the course of their evolution, have retained genes homologous to those that generate circa-12.42-hour rhythms in intertidal organisms (Zhu et al., 2023). Furthermore, the products of their expression in humans oscillate approximately twice each day. Taken together, the findings raise the possibility that an endogenous circa-12.42-hour biological tidal rhythm entrained to the 12.42-hour component of the lunisolar gravimetric cycle generates the 14.77-day semi-synodic unit of the bipolar mood cycles by beating against the 12-hour solar component of the semi-diurnal lunisolar gravimetric cycle.
Potentiation of Lunar Cycles by Antidepressant Medications
The finding in patients 2 and 4 (Figures 5g and 8a) that rapid cycling with the Moon ceased when treatment with antidepressant medications was suspended is consistent with earlier reports that such medications can induce rapid cycling (Wehr and Goodwin, 1987), and it raises the possibility that the drugs acted on a mechanism that facilitated the patients’ responses to lunar influence. If so, antidepressants and other types of drugs, could provide clues to biological mechanisms that mediate humans’ responses to lunar influence, and they could be used to probe those mechanisms. Information gained in that manner might lead to new types of treatments for bipolar disorder, and to deeper insights into the physiology and biochemistry of lunar cycles in humans.
The “Myth” of Lunar Influence
Has the myth of lunar influence on humans not been debunked? It is true that during the past half-century, studies of lunar influence and replications of those studies mostly yielded negative results. However, results of this and other longitudinal studies raise the possibility that results of the earlier studies were false negatives. Almost none of those studies examined phenomena that were known to recur in a cyclic manner, employed longitudinal designs, analyzed data on an individual case-by-case basis, and took into account the possibility that more than 1 type of lunar cycle was involved. With that approach, evidence of lunar influence could have been lost, especially when individuals’ heterogeneous responses to lunar influence were averaged together in bulk analyses (Wehr and Helfrich-Förster, 2021).
Future Directions
The results of our study pose questions for future research. For example, do lunar cycles and their mutual interactions influence the timing of manic episodes in non-circular, non-rapid cycling forms of bipolar disorder? Does the recognition of lunar influence on bipolar disorder open the way to new types of treatment based on the predictability of lunar cycles? For example, would targeted interventions that promote sleep during critical phases of lunar cycles prevent onsets of mania and thereby prevent rapid cycling (Wehr, 2018b)?
Limitations
These were retrospective analyses of mood assessments that the patients recorded prospectively in different ways. Medication treatments varied among the patients, and there were no systematic controls for effects of medications.
Although the timeseries of patients’ mood assessments spanned more than 14.5 patient-years, the number of patients was small. That number limits confidence in the validity and generalizability of the findings. Until recently, it has been difficult to obtain continuous multi-year daily records of behavioral changes in large numbers of bipolar patients because disruptions in motivation and behavior that are characteristic of the disorder can compromise the continuity of record-keeping. On the other hand, the size of the sample is not so small when viewed from a longitudinal perspective. As Kepler showed, important general principles can be extracted from a small number of cases if the periods of observation are long.
Methods for timeseries analysis, such as the Lomb-Scargle periodogram, can, by chance, produce random peaks unrelated to the process being sampled, leading to false-positive results. Such methods also can lead to false-negative results. In the present instance, in actograms and average profiles of mood cycles, relationships of mood cycles to each of the 3 types of lunar cycles appeared to be present by visual inspection of actograms and of average profiles of ratings, but were not always detected by periodogram analyses. For example, distinct peaks in average profiles of patient 3’s mania ratings coincided with syzygies of the synodic cycle, standstills of the tropical cycle, and perigees of the anomalistic cycle, and z-scores associated with peaks and troughs in all three of those profiles were statistically significant (Figure 7b). It seems unlikely that peaks in mania coincided with peaks in extreme phases of all 3 types of lunar cycles by chance alone. Nevertheless, a Lomb-Scargle periodogram detected the period of only 2 types of lunar cycle, the semi-synodic and semi-tropical cycles. Similarly, in actograms of patient 4’s mania ratings, associations of onsets of mania with syzygies of the synodic cycle, standstills of the tropical cycle, and perigees of the anomalistic appeared to be present by visual inspection of actograms (Figure 9a-9e). It seems unlikely that associations of onsets of mania with extreme phases of 3 types of lunar cycles occurred by chance alone. Nevertheless, a Lomb-Scargle periodogram detected the period of only one of the 3 types of lunar cycles, the semi-tropical cycle (Figure 3). In contrast, an eigenvalue pencil test detected periods near integer multiples of both semi-synodic and semi-tropical lunar cycles (Table 1).
In our view, visual inspection of data that have been organized in the format of actograms can provide useful information about timeseries. While visual inspection can be misleading, it can be used, as here, to verify, explore, and understand positive findings, and to question negative findings that arise from other types of analysis.
Finally, of course, it is important to acknowledge that correlation does not prove causation. At the same time, correlation adds to plausibility, and it provides a rationale for further research.
Footnotes
Acknowledgements
The authors wish to thank the patients for their substantial contributions to the study, and Elizabeth Wehr for advice and editing. The views expressed in this article do not necessarily represent the views of the NIMH or the US Federal Government.
Conflict of Interest Statement
The authors have no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
