Circadian Rhythm

A circadian rhythm is a cycle of close to 24 hours in physiology and behavior that your body generates on its own and resets every day from outside time cues, mainly light. In humans a master clock in the suprachiasmatic nucleus of the hypothalamus keeps that rhythm and coordinates the clocks in muscle, liver and most other tissues.

The rhythm explains why you feel sleepy at 3 AM and alert at 3 PM on the same amount of sleep, why your heart rate, HRV and core temperature all trace a daily curve, and why a 2 hour weekend lie-in leaves you groggy on Monday. It also explains why the timing of your sleep, and how steady that timing is from night to night, predicted mortality more strongly than sleep duration in a cohort of nearly 61,000 adults.

01The clock in the hypothalamus

The suprachiasmatic nucleus (SCN) holds about 10,000 neurons on each side of the third ventricle (Hastings et al., 2018). In 1972 two groups showed that destroying it removes daily rhythms. Rats with bilateral SCN lesions permanently lost their circadian rhythms of drinking and locomotor activity (Stephan and Zucker, 1972), and lesioned rats lost the daily rhythm of adrenal corticosterone (Moore and Eichler, 1972). Isolated in culture, the SCN keeps time on its own with high precision (Hastings et al., 2018).

Inside each clock cell, a transcription and translation feedback loop keeps the time. The CLOCK and BMAL1 proteins switch on the Period and Cryptochrome genes, whose protein products build up and shut down their own activation, and one turn of that loop takes about a day (Takahashi, 2017). Nearly every tissue runs a version of the same loop. The SCN keeps those peripheral clocks in step, and they can dissociate from it when feeding or activity is shifted (Borbély et al., 2016).

Left without time cues, the human clock runs slightly long. Under tightly controlled lighting, the free-running period averaged 24.18 hours in young and older adults alike (Czeisler et al., 1999). In a larger month-long inpatient study of 157 people, it averaged 24.15 hours, with women at 24.09 hours and men at 24.19 hours. Of those, 35% of women and 14% of men had a period shorter than 24 hours (Duffy et al., 2011). A 24.15 hour clock drifts about 9 minutes later every day unless something resets it. Daily resetting is the job of the time cues described below.

02Two processes decide when you sleep

Alexander Borbély's two-process model has framed sleep regulation since the early 1980s (Borbély, 2022). Process S is sleep pressure, a homeostatic drive that builds the longer you stay awake and discharges during sleep. Its marker is slow-wave activity in the non-REM sleep EEG, which falls exponentially across a night and rises after sleep deprivation. Process C is the circadian signal from the SCN, which raises and lowers your readiness for sleep on a fixed daily schedule regardless of how long you have been awake (Borbély et al., 2016).

You fall asleep easily when high sleep pressure coincides with the circadian window for sleep. You struggle when the two disagree. After an all-nighter, sleep pressure is at its highest by morning, yet the model predicts a second wind as the circadian signal for wakefulness rises with the day. A night-shift worker trying to sleep at 9 AM faces the reverse, with plenty of sleep pressure and a clock pushing for wakefulness. The model also explains why sleep debt and circadian misalignment are separate problems. Debt is a Process S shortfall. Misalignment is sleeping at the wrong phase of Process C, and it can happen with no debt at all.

03Time cues that reset the clock

Chronobiologists call a time cue a zeitgeber, German for "time giver." Each one shifts the clock earlier (a phase advance) or later (a phase delay) depending on when in your circadian day it arrives. A plot of shift against timing is a phase response curve.

Light

Light is the dominant zeitgeber. A class of retinal ganglion cells containing the photopigment melanopsin responds to light without input from rods or cones and projects to the SCN (Berson et al., 2002).

The direction of the shift depends on timing. In 21 adults given 6.7 hours of bright light at different circadian phases, light centered before the core body temperature minimum delayed the clock, and light centered after it advanced the clock. The full curve spanned 5.02 hours from largest delay to largest advance (Khalsa et al., 2003). The temperature minimum usually falls near the middle of a night's sleep (Baehr et al., 2000), so bright light soon after waking moves the clock earlier and bright light late in the evening moves it later.

An expert consensus published in PLOS Biology translated this into targets measured as melanopic equivalent daylight illuminance (melanopic EDI) at the eye. It recommends at least 250 lux during the day, no more than 10 lux starting at least 3 hours before bedtime, and no more than 1 lux in the sleep environment (Brown et al., 2022).

Meals

Meal timing resets the clocks in peripheral tissues and leaves the master clock largely alone. Ten young men in a 13-day lab protocol moved all three meals 5 hours later. Melatonin, cortisol and subjective sleepiness did not shift. The daily rhythm of plasma glucose shifted 5.69 hours later, and the clock gene PER2 in fat tissue shifted 0.97 hours later (Wehrens et al., 2017). Late meals can leave your metabolic clocks out of step with your brain's clock.

Exercise

Exercise shifts the human clock with a phase response curve of its own. In 99 adults who did 60 minutes of moderate treadmill exercise on three consecutive days at one of eight clock times, exercise at 7 AM and between 1 and 4 PM advanced the melatonin rhythm, and exercise between 7 and 10 PM delayed it. The size of the shifts was comparable to what bright light of the same duration would produce, with no significant differences by age or sex (Youngstedt et al., 2019). In mice, scheduled exercise shifted the molecular clock in skeletal muscle and lung while leaving the SCN in place (Wolff and Esser, 2012).

Evening training does not generally ruin sleep. A meta-analysis of 23 studies found that a single evening session did not impair sleep in healthy adults, with the exception that vigorous exercise ending 1 hour or less before bedtime might delay sleep onset and reduce total sleep time and efficiency (Stutz et al., 2019).

Caffeine

Caffeine acts on the clock as well as on sleep pressure. A dose equal to a double espresso taken 3 hours before bedtime delayed the melatonin rhythm by about 40 minutes, nearly half the delay produced by 3 hours of bright evening light (Burke et al., 2015).

04Chronotype

Your chronotype is where your clock settles relative to the solar day, early for "larks," late for "owls," and somewhere between for most people. The Horne and Östberg Morningness-Eveningness Questionnaire measures it as a preference, and morning types on that scale reach their daily temperature peak earlier than evening types (Horne and Östberg, 1976). The Munich ChronoType Questionnaire measures behavior instead. It takes the midpoint of your sleep on free days and corrects it for any sleep debt carried over from workdays (Roenneberg et al., 2004).

Chronotype changes with age. In about 25,000 people, mostly from Germany and Switzerland, children were early chronotypes who grew steadily later through adolescence. Women reached their latest timing at 19.5 years and men at 20.9 years, after which both grew earlier again. The sex difference disappeared at around 50 (Roenneberg et al., 2004). Genes matter as well. A study of 697,828 people found 351 genetic loci linked to being a morning person. The 5% of people carrying the most morningness alleles slept 25 minutes earlier on average than the 5% carrying the fewest (Jones et al., 2019).

Chronotype shows up in core temperature. In 172 young adults, the daily temperature minimum fell at 03:50 on average for morning types, 05:02 for intermediate types and 06:01 for evening types. Evening types slept on an earlier part of their temperature cycle, which the authors linked to their greater alertness at bedtime and sleepiness after waking (Baehr et al., 2000). An owl on a 6:30 AM alarm is waking close to the bottom of that curve.

05Social jet lag

Social jet lag is the gap between when your clock wants you asleep and when your schedule allows it, measured as the difference in sleep midpoint between free days and workdays (Wittmann et al., 2006). If you sleep 11:30 PM to 6:30 AM on weekdays, your midpoint is 3:00 AM. If you sleep 1:00 AM to 9:00 AM on weekends, your midpoint is 5:00 AM. The difference is 2 hours of social jet lag, the circadian equivalent of flying two time zones west every Friday and back every Monday.

Late chronotypes carry the most social jet lag, because work and school schedules cut their sleep short on workdays and they compensate on free days. In 501 volunteers, chronotype was linked to psychological wellbeing and stimulant use, most strongly in people up to age 25, and the authors attributed those links mainly to social jet lag (Wittmann et al., 2006). In a large epidemiological sample, social jet lag was associated with higher body mass index beyond what sleep duration explained (Roenneberg et al., 2012). Both studies are observational. A lab trial of weekend recovery sleep found that the recovery group's circadian timing had shifted later when weekday restriction resumed, and the sleep debt article covers its metabolic results.

06Sleep regularity and mortality

The Sleep Regularity Index (SRI) measures how likely you are to be in the same state, asleep or awake, at any two moments 24 hours apart. A score of 100 means perfectly regular sleep and wake times, and 0 means random timing (Phillips et al., 2017; Windred et al., 2024). It rewards stable timing directly. Two nights of 8 hours at different clock times score lower than two nights of 8 hours at the same clock times.

In 61 undergraduates tracked for 30 days, the least regular fifth had a dim-light melatonin onset at 00:08 on average against 21:32 in the most regular fifth, a gap of more than two and a half hours. A model of the circadian pacemaker traced most of that difference to the irregular sleepers' light exposure (Phillips et al., 2017).

Windred and colleagues (2024) calculated SRI from more than 10 million hours of wrist accelerometer data in 60,977 UK Biobank participants with a mean age of 62.8 years. The median SRI was 81.0, with an interquartile range of 73.8 to 86.3. Over a mean follow-up of 6.3 years, 1,859 participants died. Compared with the least regular fifth (SRI below about 72), the four more regular fifths had a 20% to 48% lower risk of all-cause mortality, a 16% to 39% lower risk of cancer mortality and a 22% to 57% lower risk of cardiometabolic mortality, after adjustment for age, sex, ethnicity, sociodemographic, lifestyle and health factors. Models built on regularity fit the all-cause mortality data better than equivalent models built on sleep duration. The study is observational, so it shows an association. Irregular sleep may partly mark illness, shift work or other exposures that the adjustments did not capture.

A National Sleep Foundation expert panel reviewed 63 publications and reached consensus that regular sleep timing is important for health and performance. The panel also agreed that when weekday sleep is too short, catch-up sleep on free days may be beneficial (Sletten et al., 2023). Regular timing is the target, and extra weekend sleep is a partial remedy for a short week.

07Heart rate, HRV and temperature across the day

Your cardiovascular readings follow the clock even when behavior is held constant. In a constant routine protocol that removed sleep, meals and activity as confounds, heart rate peaked in the afternoon at about 16:36 and HRV peaked in the early morning, between 05:00 and 06:59 (Vandewalle et al., 2007). During sleep, parasympathetic activity was highest in slow-wave sleep at around 02:00, and sympathetic activity during REM sleep peaked in the early morning (Boudreau et al., 2013). The sleep stages entry covers why deep sleep and REM affect the heart so differently.

This has a direct consequence for your own numbers. An HRV reading taken at 3 PM and one taken at 6 AM sit at different points on the same daily curve, and the gap between them says little about recovery. Resting heart rate behaves the same way. Readings compare best when they come from the same circadian window, which overnight measurements on a steady sleep schedule provide. Shift your bedtime by 3 hours and your overnight window samples a different part of the rhythm.

Core body temperature bottoms out during sleep, at between 03:50 and 06:01 on average across chronotypes in Baehr's sample (Baehr et al., 2000). Apple Watch Series 8 and later, all Apple Watch Ultra models and Apple Watch SE 3 sample wrist temperature every 5 seconds during sleep and report each night as a deviation from a baseline set over about 5 nights. Apple notes that diet, exercise, alcohol, the sleep environment, menstrual cycles and illness all move the reading, and the feature is not a medical device (Apple, 2026a). A nightly deviation cannot tell you your circadian phase. It is useful for spotting nights that break from your pattern.

08Training performance by time of day

Most physical performance peaks in the afternoon or evening. A meta-analysis of studies that tested people at three or more times of day found effect sizes favoring evening performance of 0.23 for endurance tests, 0.73 for Wingate peak power, 0.39 for handgrip strength and 0.79 for jump height. The authors rated the evidence strong that anaerobic power and jump height peak between 13:00 and 20:00, and found little evidence of a peak time for maximal endurance (Knaier et al., 2022).

Clock time is a rough proxy. In athletes tested across the day, time since waking predicted peak performance better than time of day, and individual performance varied by as much as 26% over a day (Facer-Childs and Brandstaetter, 2015). An owl who wakes at 9 AM and a lark who wakes at 5 AM are at different biological times at 8 AM.

Regular training at one time of day narrows the gap. A review of time-specific training found that adaptations are greater at the time of day you usually train (Chtourou and Souissi, 2012). A meta-analysis of 11 resistance training studies found that morning training raised morning strength to evening levels, and that overall gains in strength and muscle size were similar whether people trained in the morning or the evening (Grgic et al., 2019). Train at the time that fits your life, and schedule key sessions and tests near the time you will compete.

09How wearables measure circadian rhythm

No consumer wearable measures circadian phase directly. The laboratory reference is dim-light melatonin onset, which requires hourly saliva samples in dim light (Cheng et al., 2021). Watches record proxies such as sleep timing, heart rate, movement, light and skin temperature, and models estimate phase from them.

Sleep timing is the most reliable of those proxies. In a lab study of six wrist devices against polysomnography, every device detected more than 90% of sleep epochs but only 29% to 52% of wake epochs, and the Apple Watch Series 8 had the best agreement, a Cohen's kappa of 0.53 (Schyvens et al., 2025). Your watch finds the start and end of sleep well, and it may credit you with sleep during still, wakeful stretches. The Apple Watch sleep accuracy article covers the details.

Models can go further. An actigraphy and light model predicted melatonin onset in 45 night-shift workers with a concordance of 0.70 and a mean absolute error of 2.88 hours, with 76% of predictions within 2 hours (Cheng et al., 2021). Another method extracted a circadian rhythm in heart rate from more than 130,000 days of wearable data from medical interns (Bowman et al., 2021). These are research tools. Apple's own Sleep Score includes timing, with 30 of its 100 points for bedtime consistency, which considers when you fell asleep during the last 13 nights (Apple, 2026b).

10How Titan tracks sleep consistency

Titan does not estimate circadian phase. It gives you two numbers on the Sleep screen, a nightly Consistency score and a Target Wake Time, and it reads Recovery from an overnight window.

Consistency score. For each night, Titan takes the median hours asleep across the nights with sleep data in the 7 days ending that night. It needs at least 3 such nights. It then scores the night as 1 − |hours asleep − median| ÷ 2, capped between 0% and 100%. A night 30 minutes from the median scores 75%, 1 hour scores 50%, and 2 hours or more scores 0%. Longer and shorter nights are penalized equally. If your median is 7.5 hours and you sleep 6.5, that night scores 50%. The Sleep history chart shows the score over 7, 30 or 60 days.

The score is built from hours asleep. Bedtime and wake time do not enter it, so it differs from the SRI. Two 8 hour nights, one from 11 PM to 7 AM and one from 1 AM to 9 AM, both score 100% if your median is 8 hours. The SRI would mark the second as irregular.

Target Wake Time. Titan sets your target as the median time your counted sleep sessions ended over the last 14 days. With no sessions in that window it shows 7:00 AM. Comparing today's wake time with this target is the closest Titan gets to a timing measure, so read it alongside the consistency score. A steady consistency score with wake times swinging 2 hours around the target points to social jet lag that the score alone misses.

Titan Age. Sleep consistency is one of the Recovery signals in Titan Age. Titan averages your daily consistency across the week, and a higher average moves the estimate younger.

Recovery. Titan scores Recovery from the HRV and heart rate recorded during the main sleep session that ended that morning. Because the window follows your sleep, a steady schedule compares each night at a similar circadian phase. When no sleep session exists, Titan reads HRV and heart rate from 10 PM to 10 AM and marks the score Partial night. Each session counts for the day it ends on, provided it ends by noon, so daytime sleep after a night shift that runs past noon counts for no day (how sleep is detected).

The sleep consistency help article walks through the screen itself, and the sleep guide for athletes covers how regularity, debt and travel fit into training.

11Keeping your rhythm steady

Anchor your wake time first. It sets your first light exposure of the day, and a fixed wake time keeps that light at the same clock time every morning. Get bright light soon after waking, outdoors if you can, and aim for the 250 lux melanopic EDI daytime target. Dim your lights for the last 3 hours before bed.

Keep weekend wake times close to weekday ones, since every hour of difference in sleep midpoint is an hour of social jet lag. If a short week leaves you in debt, an earlier bedtime repays some of it without moving your morning light.

Time your other cues deliberately. Keep caffeine out of the evening, since a double espresso 3 hours before bed delayed melatonin by 40 minutes in the Burke study. Keep hard evening sessions from ending within an hour of bedtime. Eat your last large meal at a consistent time.

In your data, watch the Target Wake Time gap and the consistency trend over weeks. Compare HRV and resting heart rate only across nights with similar sleep timing. When you test performance, test at the same time of day and at a similar time since waking.

12References

• Apple (2026a). Track your nightly wrist temperature changes with Apple Watch. Apple Support, accessed September 2026. https://support.apple.com/en-us/102674
• Apple (2026b). View your sleep score on Apple Watch. Apple Watch User Guide, accessed September 2026. https://support.apple.com/guide/watch/view-your-sleep-score-apded441a669/watchos
• Baehr EK et al. (2000). Individual differences in the phase and amplitude of the human circadian temperature rhythm: with an emphasis on morningness-eveningness. Journal of Sleep Research 9(2):117-127. https://doi.org/10.1046/j.1365-2869.2000.00196.x
• Berson DM et al. (2002). Phototransduction by retinal ganglion cells that set the circadian clock. Science 295(5557):1070-1073. https://doi.org/10.1126/science.1067262
• Borbély AA et al. (2016). The two-process model of sleep regulation: a reappraisal. Journal of Sleep Research 25(2):131-143. https://doi.org/10.1111/jsr.12371
• Borbély A (2022). The two-process model of sleep regulation: beginnings and outlook. Journal of Sleep Research 31(4):e13598. https://doi.org/10.1111/jsr.13598
• Boudreau P et al. (2013). Circadian variation of heart rate variability across sleep stages. Sleep 36(12):1919-1928. https://doi.org/10.5665/sleep.3230
• Bowman C et al. (2021). A method for characterizing daily physiology from widely used wearables. Cell Reports Methods 1(4):100058. https://doi.org/10.1016/j.crmeth.2021.100058
• Brown TM 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
• Burke TM et al. (2015). Effects of caffeine on the human circadian clock in vivo and in vitro. Science Translational Medicine 7(305):305ra146. https://doi.org/10.1126/scitranslmed.aac5125
• Cheng P et al. (2021). Predicting circadian misalignment with wearable technology: validation of wrist-worn actigraphy and photometry in night shift workers. Sleep 44(2):zsaa180. https://doi.org/10.1093/sleep/zsaa180
• Chtourou H, Souissi N (2012). The effect of training at a specific time of day: a review. Journal of Strength and Conditioning Research 26(7):1984-2005. https://doi.org/10.1519/JSC.0b013e31825770a7
• Czeisler CA et al. (1999). Stability, precision, and near-24-hour period of the human circadian pacemaker. Science 284(5423):2177-2181. https://doi.org/10.1126/science.284.5423.2177
• Duffy JF et al. (2011). Sex difference in the near-24-hour intrinsic period of the human circadian timing system. Proceedings of the National Academy of Sciences 108(Suppl 3):15602-15608. https://doi.org/10.1073/pnas.1010666108
• Facer-Childs E, Brandstaetter R (2015). The impact of circadian phenotype and time since awakening on diurnal performance in athletes. Current Biology 25(4):518-522. https://doi.org/10.1016/j.cub.2014.12.036
• Grgic J et al. (2019). The effects of time of day-specific resistance training on adaptations in skeletal muscle hypertrophy and muscle strength: a systematic review and meta-analysis. Chronobiology International 36(4):449-460. https://doi.org/10.1080/07420528.2019.1567524
• Hastings MH et al. (2018). Generation of circadian rhythms in the suprachiasmatic nucleus. Nature Reviews Neuroscience 19(8):453-469. https://doi.org/10.1038/s41583-018-0026-z
• Horne JA, Östberg O (1976). A self-assessment questionnaire to determine morningness-eveningness in human circadian rhythms. International Journal of Chronobiology 4(2):97-110.
• Jones SE et al. (2019). Genome-wide association analyses of chronotype in 697,828 individuals provides insights into circadian rhythms. Nature Communications 10(1):343. https://doi.org/10.1038/s41467-018-08259-7
• Khalsa SBS et al. (2003). A phase response curve to single bright light pulses in human subjects. Journal of Physiology 549(3):945-952. https://doi.org/10.1113/jphysiol.2003.040477
• Knaier R et al. (2022). Diurnal variation in maximum endurance and maximum strength performance: a systematic review and meta-analysis. Medicine and Science in Sports and Exercise 54(1):169-180. https://doi.org/10.1249/MSS.0000000000002773
• Moore RY, Eichler VB (1972). Loss of a circadian adrenal corticosterone rhythm following suprachiasmatic lesions in the rat. Brain Research 42(1):201-206. https://doi.org/10.1016/0006-8993(72)90054-6
• Phillips AJK et al. (2017). Irregular sleep/wake patterns are associated with poorer academic performance and delayed circadian and sleep/wake timing. Scientific Reports 7(1):3216. https://doi.org/10.1038/s41598-017-03171-4
• Roenneberg T et al. (2004). A marker for the end of adolescence. Current Biology 14(24):R1038-R1039. https://doi.org/10.1016/j.cub.2004.11.039
• Roenneberg T et al. (2012). Social jetlag and obesity. Current Biology 22(10):939-943. https://doi.org/10.1016/j.cub.2012.03.038
• Schyvens AM et al. (2025). A performance validation of six commercial wrist-worn wearable sleep-tracking devices for sleep stage scoring compared to polysomnography. Sleep Advances 6(2):zpaf021. https://doi.org/10.1093/sleepadvances/zpaf021
• Sletten TL et al. (2023). The importance of sleep regularity: a consensus statement of the National Sleep Foundation sleep timing and variability panel. Sleep Health 9(6):801-820. https://doi.org/10.1016/j.sleh.2023.07.016
• Stephan FK, Zucker I (1972). Circadian rhythms in drinking behavior and locomotor activity of rats are eliminated by hypothalamic lesions. Proceedings of the National Academy of Sciences 69(6):1583-1586. https://doi.org/10.1073/pnas.69.6.1583
• Stutz J et al. (2019). Effects of evening exercise on sleep in healthy participants: a systematic review and meta-analysis. Sports Medicine 49(2):269-287. https://doi.org/10.1007/s40279-018-1015-0
• Takahashi JS (2017). Transcriptional architecture of the mammalian circadian clock. Nature Reviews Genetics 18(3):164-179. https://doi.org/10.1038/nrg.2016.150
• Vandewalle G et al. (2007). Robust circadian rhythm in heart rate and its variability: influence of exogenous melatonin and photoperiod. Journal of Sleep Research 16(2):148-155. https://doi.org/10.1111/j.1365-2869.2007.00581.x
• Wehrens SMT et al. (2017). Meal timing regulates the human circadian system. Current Biology 27(12):1768-1775. https://doi.org/10.1016/j.cub.2017.04.059
• Windred DP et al. (2024). Sleep regularity is a stronger predictor of mortality risk than sleep duration: a prospective cohort study. Sleep 47(1):zsad253. https://doi.org/10.1093/sleep/zsad253
• Wittmann M et al. (2006). Social jetlag: misalignment of biological and social time. Chronobiology International 23(1-2):497-509. https://doi.org/10.1080/07420520500545979
• Wolff G, Esser KA (2012). Scheduled exercise phase shifts the circadian clock in skeletal muscle. Medicine and Science in Sports and Exercise 44(9):1663-1670. https://doi.org/10.1249/MSS.0b013e318255cf4c
• Youngstedt SD et al. (2019). Human circadian phase-response curves for exercise. Journal of Physiology 597(8):2253-2268. https://doi.org/10.1113/JP276943