Multiday rhythms modulate human heart rate: an observational study in healthy adults

This observational study of 623 healthy adults utilizing multi-year wearable data reveals that significant multiday (infradian) heart rate rhythms are prevalent in the general population, manifesting as distinct chronotypes influenced by environmental cycles and menstrual phases, thereby highlighting the need to investigate their mechanisms and clinical relevance beyond traditional circadian research.

The Gist

Imagine your body isn't just a machine that runs on a 24-hour clock (like a daily alarm). Instead, it's more like a grand orchestra playing a symphony that has many different rhythms happening at once.

Most people know about your circadian rhythm—the 24-hour cycle that tells you when to wake up and when to sleep, driven by the sun. But this new study discovered that your heart has a whole other set of "internal clocks" that tick much slower: rhythms that last for weeks, months, or even a year.

Here is the story of what the researchers found, explained simply.

1. The Hidden "Heartbeat" of the Year

The researchers looked at data from over 600 healthy college students who wore smartwatches for up to four years. They were looking for patterns in their heart rates that weren't just daily.

The Discovery:
They found that 7 out of 10 people had a "secret rhythm" in their heart rate that lasted longer than a week.

• Some hearts beat in a weekly pattern (like a 7-day cycle).
• Some had a monthly pattern (around 25 or 35 days).
• Some had seasonal patterns (like a 6-month or 12-month cycle).

The Analogy:
Think of your heart rate like a tide. You know the tides go up and down every day (circadian). But this study found that there are also "super-tides" that rise and fall over weeks or months, regardless of whether the sun is shining or the moon is full.

2. It's Not Just About the Weather or the Calendar

You might think, "Oh, maybe our hearts speed up in winter because it's cold, or slow down on weekends because we party."

The researchers checked this.

• The Weekend Effect: Yes, heart rates were slightly higher on Fridays and Saturdays.
• The Seasonal Effect: Yes, heart rates were slightly lower in summer.
• The Moon Effect: No, the moon didn't seem to matter.

The Twist:
Even after they mathematically "subtracted" the effects of weekends, seasons, and the weather, the long rhythms were still there.
This means your heart has its own internal metronome that keeps ticking even when the outside world changes. It's like a drummer in a band who keeps a steady beat even if the singer stops singing or the lights go out.

3. The "Chronotypes" (Your Heart's Personality)

Just as some people are "night owls" and others are "early birds" for sleep, this study found that people have different Heart Rhythm Personalities (called chronotypes):

• The Weekly Heart: Mostly found in men. Their heart rate seems to cycle every 7 days.
• The Short-Monthly Heart: Mostly found in women. Their heart rate cycles every 25 days.
• The Long-Monthly Heart: Found in both men and women, cycling every 35 days.
• The Multi-Month Heart: Mostly found in men, with cycles lasting over 35 days.

The Analogy:
Imagine a dance floor. Some people dance to a fast 4/4 beat (weekly), some to a slow waltz (monthly), and some to a very slow, deep bass line (yearly). The study found that your gender and lifestyle help decide which "dance" your heart prefers.

4. The Hormone Connection (The "Female" Rhythm)

The researchers also looked at women who tracked their menstrual cycles.

• The Sync: For most women, their heart rate rhythm was perfectly synchronized with their period. Their heart rate would peak right before their period started and drop during the "follicular" phase (after the period).
• The Surprise: But here's the kicker: Men and older women (who have stopped menstruating) also had monthly heart rhythms.

The Analogy:
Think of the menstrual cycle as a conductor for the orchestra in women. It tells the heart, "Play loud now, play soft later." But for men and older women, the conductor is missing, yet the music (the monthly rhythm) is still playing. This suggests that the monthly rhythm is a fundamental part of human biology, not just a side effect of female hormones.

5. Friends Influence Your Heartbeat

The study looked at social networks. They found that close friends, roommates, and siblings had heart rhythms that were more similar to each other than strangers.

The Analogy:
If you live with a roommate, you might start eating lunch at the same time or sleeping at the same time. This study suggests you might also be "syncing" your internal biological clocks. It's like two tuning forks: if you strike one, the other starts vibrating at the same frequency. Your friends might be subtly tuning your heart's rhythm to match theirs.

Why Does This Matter?

This is a big deal for three reasons:

1. It's Normal: We used to think heart rate was just about "now" (exercise, stress, sleep). Now we know it has a long-term "weather forecast" built into it.
2. Better Health Tracking: If you have a condition like epilepsy, depression, or heart disease, your symptoms might flare up on a specific part of your monthly or yearly rhythm. Knowing your "Heart Rhythm Personality" could help doctors predict when you might feel sick and treat you before it happens.
3. Women's Health: It offers a new way to track hormonal health. Instead of just guessing when your period is coming, your smartwatch could detect the subtle heart rate shifts that happen days before, giving you a more accurate picture of your body.

The Bottom Line

Your body is a complex, multi-layered machine. While your daily sleep-wake cycle is the most obvious rhythm, your heart is also dancing to a much slower, deeper beat that lasts for weeks and months. This rhythm is part of who you are, influenced by your gender, your friends, and your own internal biology, and it's been there all along, waiting for us to finally listen.

Technical Summary

1. Problem Statement

Chronobiology has historically focused on circadian rhythms (approx. 24-hour cycles). However, infradian rhythms (cycles longer than 24 hours, ranging from days to months) are prevalent in human physiology but remain poorly understood due to a lack of longitudinal datasets and robust detection methods.

• Knowledge Gap: While isolated studies have identified infradian rhythms in specific clinical populations (e.g., epilepsy, cardiovascular disease) or specific contexts (e.g., menstrual cycles), there is no systematic characterization of these rhythms in the general, healthy population.
• Uncertainty: It is unclear whether these rhythms are driven primarily by external environmental cues (zeitgebers like day-of-week, lunar phases, seasons) or by intrinsic, endogenous biological mechanisms.
• Objective: To investigate the prevalence, heterogeneity, and drivers of multiday (infradian) heart rate rhythms in a large cohort of healthy young adults using continuous wearable data.

2. Methodology

The study employed a multi-cohort approach utilizing advanced signal processing and statistical modeling.

A. Datasets

1. Primary Cohort (NetHealth): 623 university students (USA) wearing Fitbit smartwatches for up to 4 years (2015–2019).
   • Inclusion Criteria: Continuous heart rate data for $\ge$ 3 months with >80% adherence.
   • Final Sample: 523 participants (263 female, 317 male).
   • Data: Daily mean heart rate, physical activity minutes, sleep duration, and survey data (circadian preference, social networks).
2. Secondary Cohort (Menstrual): 70 participants (47 menstruating females, 11 menopausal females, 12 males) from a study in China, wearing Huawei/Fitbit devices with minute-level heart rate data and self-tracked menstrual/ovulation logs.

B. Data Preprocessing

• Detrending: To isolate intrinsic rhythms from activity-induced heart rate spikes, daily physical activity was regressed out of the daily mean heart rate using a linear model ($Y_i(n) = r_i \cdot x_i(n) + c_i$).
• Standardization: Data segments were z-scored (mean subtracted, divided by SD) to allow cross-subject comparison.
• Imputation: Missing data (<20% of segment) was interpolated using global averages.

C. Signal Analysis & Rhythm Detection

• Wavelet Analysis: A Morlet wavelet transform was applied to generate global wavelet power spectra (periodograms).
• Significance Testing: Peaks crossing the 95% significance threshold ($\alpha > 0.05$) were identified as significant rhythms.
• Period Range: Analyzed cycles from 4 days up to one-third of the total recording length (minimum 3 cycles observed).
• Clustering (Chronotypes): Unsupervised Non-Negative Matrix Factorization (NMF) was used to cluster individuals based on the distribution of their wavelet power spectra, identifying dominant "multiday chronotypes."

D. Statistical Modeling

• Environmental Regression: Generalized Additive Models (GAM) were used to regress out day-of-week, lunar, and seasonal effects to test for endogenous drivers.
• Demographic Analysis: Multinomial logistic regression assessed the association between chronotypes and sex, circadian preference, activity levels, and sleep duration.
• Synchrony Analysis: Mann-Whitney U tests compared wavelet periodogram similarity between socially "close" pairs (roommates/friends) vs. "distant" pairs.

3. Key Contributions

1. First Large-Scale Characterization: This is the first study to systematically map infradian heart rate rhythms across a large, healthy population (N=523) rather than just clinical cohorts.
2. Discovery of "Multiday Chronotypes": Identified four distinct clusters of heart rate rhythms based on dominant periodicities:
   • Weekly: ~7 days (predominantly male).
   • Shorter-Monthly: ~25 days (predominantly female).
   • Longer-Monthly: ~35 days (mixed sex).
   • Multi-Month: >35 days (predominantly male).
3. Decoupling from Environment: Demonstrated that while environmental factors (weekends, seasons) influence heart rate, the specific periods of infradian rhythms are often not tightly locked to external cues (e.g., weekly rhythms averaged 6.3 days, not exactly 7).
4. Hormonal Coupling: Provided evidence of strong synchronization between heart rate rhythms and the menstrual cycle in females, while also confirming the existence of monthly rhythms in males and menopausal women, suggesting non-hormonal drivers for monthly cycles.

4. Key Results

Prevalence and Periodicity

• 69.7% of participants had at least one significant multiday rhythm.
• 35.9% had two or more distinct rhythms.
• Most Common Cycles:
  • Biannual (5–7 months): 47.8% of the cohort (most common dominant cycle: 31.9%).
  • Annual (11–13 months): 38.9% (dominant: 24.2%).
  • 10-Week (8–12 weeks): 32.8% (dominant: 18.9%).
  • Weekly (5–9 days): 15.7% (dominant: 7.1%).
  • Monthly (25–35 days): 9.6% (dominant: 5.0%).

Environmental vs. Endogenous Drivers

• Day-of-Week: A "weekend effect" (higher HR on Fri/Sat) was observed. However, when regressed out, the prevalence of precise 7-day rhythms dropped significantly, suggesting many "weekly" rhythms are behavioral.
• Seasonality: HR was lower in Spring/Summer and higher in Winter/Autumn.
• Lunar Cycle: No significant effect was found.
• Endogeneity: The most common rhythms (biannual, 10-week) lack clear environmental explanations, suggesting intrinsic biological oscillators.

Demographic and Behavioral Correlates

• Sex: The strongest predictor of chronotype. Females were significantly more likely to have monthly rhythms; males were more likely to have weekly and multi-month rhythms.
• Activity/Sleep: Less active individuals showed patterns similar to females (monthly rhythms). Longer sleep duration was associated with shorter monthly rhythms (though not statistically significant in all models).
• Social Synchrony: Individuals in close social relationships (roommates, friends) exhibited significantly higher similarity in their wavelet periodogram distributions compared to strangers ($p=0.004$).

Menstrual Synchronization (Secondary Cohort)

• 75.7% of this smaller cohort had significant multiday rhythms.
• Menstruating Females: 48.9% had monthly rhythms. Resting HR peaked during the luteal phase and was lowest during the follicular phase.
• Synchronization: Menstruation was strongly synchronized to the peak of the heart rate rhythm in most menstruating females (Median Synchronization Index: 0.75).
• Non-Females: Monthly rhythms were also present in males (16.7%) and menopausal women (27.3%), indicating these rhythms are not solely driven by reproductive hormones.

5. Significance and Implications

• Physiological Insight: The findings suggest that the human autonomic nervous system possesses a complex, multi-layered temporal architecture governed by intrinsic mechanisms that operate on timescales of weeks to months, independent of the 24-hour circadian clock.
• Clinical Relevance:
  • Disease Monitoring: Understanding normative infradian rhythms could improve the detection of abnormalities in cardiovascular disease, epilepsy, and psychiatric disorders (e.g., bipolar disorder, seasonal affective disorder) where symptoms fluctuate on multiday cycles.
  • Chronotherapy: Treatment timing could be optimized based on individual multiday cycles rather than just circadian timing.
• Women's Health: Heart rate rhythms could serve as a passive, continuous proxy for tracking hormonal and menstrual cycles, offering higher accuracy than self-reporting, especially for irregular cycles.
• Future Directions: The study highlights the need for larger, more diverse cohorts (different ages, geographies) and higher-resolution data (minute-by-minute) to further elucidate the genetic vs. environmental drivers of these rhythms.

In conclusion, this study establishes that multiday heart rate rhythms are a fundamental, widespread feature of human physiology, driven by a mix of endogenous biological clocks and social/environmental entrainment, with significant implications for personalized medicine and health monitoring.