Signs of Aging Not Visible Through Step Count Alone: New Research on Activity Timing and Internal Rhythms

Two people each walk 8,000 steps a day. One starts moving in the morning, walks intermittently during the day, and rests quietly at night. The other barely moves until near noon and accumulates steps from evening to midnight. If you only look at the total step count, their level of physical activity seems the same. However, the "time of day information" their bodies receive may not be identical.

In recent health management, metrics like step count, exercise time, and calories burned have been emphasized. However, a study published in Nature Communications in August 2026 sheds light on another aspect that cannot be captured by quantity alone. It focuses on how activities and rest are distributed over 24 hours, how distinct the waves are, and how stable they are day by day.

To summarize the research findings in one sentence: "The aging of the body was related not only to how much one moves but also to the rhythm of movement." However, it's important not to jump to conclusions. The study did not prove that "becoming a morning person rejuvenates you" or that "night owls necessarily age faster." Understanding this boundary is key to correctly interpreting this news as health information.

Tracking the "Life Waveform" of 2,222 People with Long-term Data

The study was conducted by Jinjoo Shim and Jukka-Pekka Onnela from the Harvard T.H. Chan School of Public Health. They participated in the U.S. large-scale research platform "All of Us Research Program," targeting 2,222 people who provided both Fitbit records and electronic health records. The analyzed activity data amounted to a total of 8,447 person-years, with an average age of about 60.6 years among the subjects, 68.5% of whom were women.

Traditional activity research often involved having participants wear an accelerometer for about a week and estimating lifestyle patterns from that short-term record. However, a week can include random events like colds, travel, bad weather, or busy periods. The strength of this study lies in using multi-year records obtained from commercially available wearable devices, focusing more on sustained trends rather than temporary disruptions.

The researchers did not simply compare daily step counts. They quantified the average activity level over 24 hours, the height of activity peaks, the amount of exercise during the most active continuous 10 hours, the amount of activity during the quietest continuous 5 hours, the difference between day and night, the timing of activity peaks, and the regularity across days. Furthermore, they statistically decomposed the step count curve for each time, extracting overall activity intensity, start time of movement, peak position, and patterns where peaks split between morning and evening.

In other words, they read step counts not as a single total value but as a "waveform" akin to an electrocardiogram.

What Did "Biological Age" Measure?

The comparison was made against an index called "PhenoAge." This is neither an appearance age nor an epigenetic clock that directly measures DNA changes. It combines chronological age with nine clinical blood indicators such as albumin, creatinine, blood glucose, CRP, and white blood cell count, expressing it as an age corresponding to mortality risk.

Even at the same chronological age of 60, some people have a PhenoAge equivalent to 55, while others have one equivalent to 65. The study examined the relationship between "accelerated aging" and activity rhythms by using how much PhenoAge leads or lags behind chronological age. At baseline, 1,021 people, or 45.9% of the total, were classified into the accelerated aging group.

While PhenoAge has been validated for its association with future diseases, physical function, and mortality risk, it is not the "true age" that fully predicts human aging. It is merely a statistical indicator aggregating multiple blood test values. This distinction is important because the study does not directly measure lifespan with a smartwatch but rather finds overlaps between daily activity patterns and aging indicators calculated from blood markers.

The Strongest Factor Was "Active During the Day, Resting at Night"

The analysis consistently showed that the strength of activity rhythms was significant. People with high activity levels throughout 24 hours, clear activity peaks, high activity during the most active 10 hours, and a clear distinction between active and rest periods had lower odds of being classified into the accelerated aging group.

In long-term analysis, moving one quartile higher in the index corresponded to about 26-46% lower odds of accelerated aging, depending on the index. Specifically, average activity level, activity amplitude, relative day-night difference, and exercise volume during the most active 10 hours showed protective directions. This does not merely mean "exercising intensely at once." It is more accurate to consider that the entire rhythm, where activities are concentrated during the day and decrease at night, was related.

In fact, in the accelerated aging group, the daytime activity peak was lower, and there was a tendency for more activity between midnight and 6 a.m. compared to the non-accelerated group. The baseline average step count was about 7,215 steps for the accelerated aging group and about 8,141 steps for the non-accelerated group. However, even after researchers considered daily step count and sleep time, the main associations of rhythm indicators remained. There was something that could not be explained by simply stating, "People who walk a lot are younger."

A Late Activity Peak Increased Odds of Accelerated Aging by 22%

Another focus was the timing of the activity peak. In the long-term model, moving one quartile towards a later peak time corresponded to a 22.1% higher odds of accelerated aging. If you only look at this, it might be tempting to write a sensational headline like "Night Owls Age Faster."

However, there are at least three cautions needed for this interpretation.

First, 22.1% does not mean an individual's aging speed increased by 22.1%. It is a relative difference in "odds" in a statistical model, different from an absolute increase in incidence probability.

Second, the activity peak is not a direct measurement of the body clock itself. It is the timing of behavior recorded by Fitbit, influenced by work, commuting, childcare, caregiving, illness, living environment, seasons, mood, etc. It is not a study that confirmed "physiological night owls" by measuring melatonin secretion or core body temperature.

Third, there is the possibility that a late activity peak did not accelerate aging, but rather deteriorating health prevented daytime activity, shifting life later. "Reverse causality," where cause and effect are reversed, cannot be completely ruled out.

Therefore, this result is not a verdict against night owls. It suggests that later activity times could be clues to health changes.

For Women, "Regularity" Was More Prominent

In gender-separated analyses, the strength of activity rhythms showed protective directions for both men and women. However, the association with the timing of activity and daily regularity was stronger in women, with high regularity associated with about 15% lower odds of accelerated aging. Late activity or rest onset was linked to 12-18% higher odds in women.

For men, the same relationship was not as clear, and instead, an unstable pattern with activity peaks split between early morning and after evening, and increased nighttime movement, was associated with accelerated aging. However, the reasons for gender differences have not been clarified. Multiple factors such as hormones, social roles, work patterns, disease composition, and wearable usage are possible. Although there were differences in some effect sizes by gender, not all interactions were statistically significant, so simplifications like "only women should be regular" should be avoided.

On SNS and Reader Comments, Expectations and Caution Intersect

This study was shared on social media, with future-oriented interpretations like "the shape of the day might reflect body aging" and "smartwatches hold more information than just step counts." On LinkedIn, there were reactions evaluating the idea of connecting activity indicators with clinical biomarkers as useful. On Facebook, public posts introducing the study results along with the scale of 2,222 people and 8,447 person-years were confirmed.

On the other hand, in public comments attached to the WELT article, skepticism was more noticeable than enthusiasm. Among the 35 comments visible at the time of article acquisition, mainly four points were repeatedly raised.

The first is the criticism that "correlation is not causation." There was dissatisfaction with headlines suggesting that lifestyle rhythms influence aging, while the text states that causality is unknown. In response, some defended that science gradually reduces uncertainty and that finding correlations also has value.

The second is doubt about sample bias. People who wore a Fitbit long-term and agreed to share data might have higher health awareness and economic status than the general population. In fact, this group was 68.5% women, 82.8% white, and 93.8% had a college education or higher, not directly representing the entire U.S. population.

The third is the treatment of chronotypes like morningness-eveningness. There is a concern about whether "people who are naturally night owls are considered unhealthy" and whether activity times should be compared based on elapsed time after waking rather than clock time. This is a pertinent question in interpreting the study. Step count curves are results of behavior and cannot completely separate endogenous body clocks from social time constraints.

The fourth is the irony that "it's obvious that moving during the day and resting at night is healthier." However, the value of the study lies in quantifying hypotheses that seem common sense with real-life data and blood indicators over several years. What sounds obvious is not the same as what withstands measurement.

Note that these are qualitative summaries of publicly visible posts and comments, not a public opinion survey. They do not indicate the approval-disapproval ratio across all social media, and the number of reactions changes over time. Nonetheless, they provide material to understand where people have expectations and where they feel discomfort.

Strengths of the Study and Limitations Not to Overlook

The greatest strength is that it dealt with a long timeline of daily life rather than a single point in a doctor's office. Even if disrupted for a few days due to illness or travel, months or years of records make it easier to distinguish personal patterns and sustained changes. The connection of different types of data, such as blood tests and wearables, and repeatedly tracking changes in the same person is also significant.

However, as an observational study, the effects of interventions are unknown. Researchers statistically considered many factors such as age, gender, race, BMI, income, employment, smoking, comorbidities, medication, step count, and sleep, but unmeasured factors like night shifts, childcare, meal times, light exposure, individual chronotypes, and situations where wearables are removed remain.

There are also limitations unique to commercial devices. Fitbit's step and sleep estimates depend on proprietary algorithms, and while behavior is captured, it does not directly measure the central clock in the brain or molecular clocks in each organ. Missing data can occur during charging or non-wearing. If people in poor health stop wearing the device, the way data is missing itself could bias results.

Furthermore, blood glucose and CRP used in PhenoAge are influenced by infections, chronic diseases, medications, weight, and diet. There remains the possibility that activity rhythms and PhenoAge reflect the same underlying factors. The achievement of this study should be positioned as a "candidate digital indicator for future verification" rather than a "prescription to delay aging."

What Should We Change Starting Tomorrow?

Based solely on this study, there is no need to stop exercising at night or forcibly change to a morning type. For those who can only exercise at night due to work or health, moving at a time they can continue is more in line with current public health evidence than not exercising at all. The World Health Organization recommends adults engage in 150-300 minutes of moderate-intensity aerobic exercise or 75-150 minutes of high-intensity exercise per week. This study does not replace that "quantity" foundation.

If incorporating it into real life, the following moderate uses can be considered.

• Look at whether activities are concentrated during the day and calm at night over several weeks, not just daily step counts.

• Avoid extreme fluctuations in wake-up, meal, exercise, and bedtime, including on holidays.

• When nighttime activity increases, instead of blaming the numbers, check the background such as sleep disruption, pain, stress, caregiving, or night shifts.

• Do not aim for early rising itself as a goal; create boundaries between active and rest periods within your chronotype and living conditions.

• Do not take a wearable's single-day assessment as a diagnosis; consult a healthcare provider if there are sustained changes or symptoms.

What is particularly important is not to make rhythm a "score of discipline." Shift workers, people raising infants, caregivers, and those with chronic illnesses may not be able to adhere to an ideal timetable. A disrupted graph may not be proof of weak will but a sign visualizing life burdens or health changes.

Will Smartwatches Become "Aging Diagnostic Devices"?

In the future, wearables may evolve from devices that notify step goals to devices that detect changes from one's normal state. Gradually lowering daytime activity peaks, increasing nighttime movements, and varying daily times. Combining such changes with blood tests and interviews could lead to early investigation of frailty, metabolic abnormalities, sleep disorders, mood changes, etc.

However, privacy and overdiagnosis issues are inherent. Minute-by-minute behavioral data is health information and a record of life itself, including work, outings, and sleep. It is essential to design who owns the data, ensure it is not used for insurance or employment, and prevent anxiety from false detections.

The most intriguing point shown by this study is that signs of aging may not only appear in cells or blood but also in the "shape of how we spend our day." The question is not a binary choice of being a morning or night person. It is whether you are moving enough, resting properly, and maintaining your switch in your own way. And how that pattern changes over a long time.

Step counts are easy to understand. However, health cannot be discussed by total values alone. Future wearables are likely to reflect not only how many steps we take but also what kind of rhythm our day marks.

Source URL

• WELT: Refer to the research overview and public reactions in the WELT reader comment section.
  https://www.welt.de/wissenschaft/article6a97dd709452c52238230d11/longevity-was-ihr-tagesrhythmus-darueber-verraet-wie-schnell-ihr-koerper-altert.html

• Nature Communications: Peer-reviewed research paper by Shim and Onnela. Used to confirm subjects, analysis methods, main results,