Actigraphy-Based Movement Profiles and Their Association with Circadian Rhythms Integrity in Real-World Settings

In a study of 169 young adults under real-world conditions, actigraphy-based clustering revealed that individuals with more active movement profiles exhibit significantly stronger circadian rhythm integrity and earlier circadian phases compared to less active peers, highlighting the utility of actigraphy as a multidimensional tool for assessing the link between movement behavior and circadian health.

The Gist

Imagine your body has an internal conductor, like a maestro leading an orchestra. This conductor is your circadian rhythm, the 24-hour cycle that tells your body when to wake up, when to sleep, when to digest food, and when to repair itself. When this conductor is strong and steady, the orchestra plays beautifully, and you feel healthy. When the conductor is shaky or confused, the music becomes chaotic, leading to poor health.

This study is like a detective story where researchers tried to figure out what makes that internal conductor play a better tune. They focused on one specific question: Does how much and how we move during the day affect how well our internal clock works?

Here is the breakdown of their findings, explained simply:

1. The Setup: Watching People Move

The researchers didn't put people in a lab. Instead, they gave 169 young adults (ages 18–30) a smartwatch-like device called an actigraph to wear on their wrists for about two weeks. Think of this device as a "movement spy" that recorded every step, stretch, and shuffle the participants made in their real lives.

They also tracked how much light these people saw and how well they slept.

2. The Discovery: Two Types of Movers

Using a computer program (a type of artificial intelligence), the researchers sorted everyone into two distinct groups based on their movement patterns:

• The "Less Active" (LA) Group: These people tended to sit or lie down for long stretches, get up for short bursts, and then sit down again. Their movement was like a flickering light—on and off, fragmented, and low energy.
• The "More Active" (MA) Group: These people moved more often. They didn't just sit for hours; they had longer periods of activity and shorter periods of sitting. Their movement was like a steady, flowing river rather than a series of puddles.

Crucially, these two groups were identical in age, gender, and even their natural "night owl" or "early bird" tendencies. The only difference was how they moved.

3. The Big Finding: Movement = A Stronger Conductor

The researchers measured the "integrity" of the participants' internal clocks using a score called the Circadian Function Index (CFI). Think of this score as a "health meter" for your body's rhythm, ranging from 0 (broken rhythm) to 1 (perfect rhythm).

• The Result: The "More Active" group had a significantly higher health meter score (0.81) compared to the "Less Active" group (0.69).
• The Analogy: Imagine the "Less Active" group's internal clock is a wobbly table leg; it works, but it's shaky. The "More Active" group's clock is a solid, sturdy table leg. Their bodies were better at keeping a consistent, strong rhythm of day and night.

4. The Surprising Twist: It's Not Just About Light

We often hear that sunlight is the most important thing for fixing your sleep schedule. While light is important, this study found something interesting:

• The "More Active" group had a stronger internal rhythm even though their light exposure wasn't drastically different from the "Less Active" group.
• However, the timing of their movement mattered. The "More Active" group tended to move more in the early evening. This specific timing seemed to help "lock in" their internal clock, making it more robust.

5. Sleep vs. Rhythm: Two Different Things

You might think that if your internal clock is better, you sleep better. But here is the twist:

• The study found no difference in sleep quality between the two groups. Both groups slept about the same amount and woke up about the same number of times.
• The Takeaway: You can have a "wobbly" internal clock (low CFI) but still sleep for 8 hours. Conversely, you can have a "rock-solid" internal clock (high CFI) and still sleep for 8 hours.
• The Metaphor: Think of sleep as the fuel in your car's tank, and the circadian rhythm as the engine's timing. You can have a full tank of gas (good sleep), but if the engine timing is off (weak rhythm), the car won't run efficiently. The "More Active" group had better engine timing, even if the fuel tank was the same size as the other group.

Why Does This Matter?

This study suggests that how you move is a powerful tool for tuning your body's internal clock.

• For Health: A strong internal rhythm is linked to lower risks of heart disease, diabetes, and even cancer.
• For Daily Life: You don't need to be a marathon runner. The study suggests that simply breaking up long periods of sitting with more frequent, consistent movement—especially in the late afternoon and evening—can act like a "tuning fork" for your body, making your internal clock stronger and more resilient.

In short: If you want your body's internal conductor to lead the orchestra perfectly, don't just worry about when you sleep or how much sun you get. Get moving, and keep that movement flowing throughout the day. It's a free, natural way to upgrade your body's operating system.

Technical Summary

1. Problem Statement

While the health benefits of physical activity and robust circadian rhythms are well-established, the specific mechanisms linking movement behavior patterns to circadian integrity in real-world settings remain partially understood. Current guidelines lack specific recommendations on the timing of exercise for circadian health. Furthermore, while wearable devices (actigraphy) are increasingly used to monitor activity, there is a gap in understanding how distinct, unsupervised movement profiles (derived from multidimensional accelerometer data) correlate with the overall integrity of circadian rhythms, specifically when measured by the Circadian Function Index (CFI).

2. Methodology

Study Population and Design

• Participants: A total of 169 young adults (aged 18–30) from Uruguay. The sample was a composite of newly recruited university students ($n=95$) and participants from three previous studies involving university students and dancers ($n=74$).
• Setting: Real-world conditions (free-living) during peri-equinox periods.
• Exclusion Criteria: No prior psychiatric/neurological/sleep diagnoses or use of sleep-affecting medications.

Data Collection

• Device: GeneActive Original tri-axial accelerometers worn on the non-dominant wrist for 7–24 days.
• Sampling: Recorded at 10 Hz; processed using the GGIR R package.
• Metrics: Activity estimated via Euclidean Norm Minus One (ENMO) in milligravity (mg). Light exposure and sleep parameters were also recorded.

Feature Engineering and Clustering

• Movement Features: 12 accelerometer-derived features were selected to capture five dimensions: overall activity level, total duration, frequency, bout duration, and intensity distribution (gradient/intercept). These were calculated across three intensity levels: Inactive (IN), Light Physical Activity (LPA), and Moderate-to-Vigorous Physical Activity (MVPA).
• Clustering Algorithm: K-means clustering (Hartigan-Wong algorithm) was applied to the standardized movement features to identify distinct behavioral profiles. The optimal number of clusters was determined using the NbClust function (elbow method, gap statistic), and cluster quality was validated via silhouette analysis.

Circadian and Sleep Metrics

• Circadian Integrity: Quantified using the Circadian Function Index (CFI), a composite score (0–1) integrating:
  • Interdaily Stability (IS): Day-to-day regularity.
  • Intradaily Variability (IV): Within-day fragmentation.
  • Relative Amplitude (RA): Consolidation of day vs. night activity.
• Circadian Phase: Estimated via non-parametric proxies (Midpoint of the 5 least active hours, L5c; Midpoint of the 10 most active hours, M10c) and parametric Cosinor analysis (Acrophase).
• Sleep Quality: Assessed via Sleep Duration, Sleep Efficiency (SE), Wake After Sleep Onset (WASO), and Sleep Regularity Index (SRI).

Statistical Analysis

• Comparisons: Wilcoxon rank-sum tests compared profiles (MA vs. LA).
• Associations: Mixed-effects regression models and linear regression were used to examine associations between movement profiles, time-of-day activity/light exposure, and circadian parameters (CFI, L5c).
• Correction: Multiple comparisons were adjusted using the Holm–Bonferroni method.

3. Key Results

Identification of Movement Profiles
The unsupervised algorithm identified two distinct, non-overlapping clusters:

1. More Active (MA) Profile ($n=96$): Characterized by higher average daily acceleration, significantly more time in LPA and MVPA, fewer minutes of inactivity, and more frequent/larger MVPA bouts.
2. Less Active (LA) Profile ($n=73$): Characterized by lower intensity, longer inactivity bouts, and lower frequency of activity bouts.

• Note: There were no significant differences in sex, age, BMI, or chronotype (MSFsc) between the two groups, suggesting the profiles are driven by behavior rather than demographics.

Circadian Rhythm Integrity (CFI)

• Primary Finding: The MA profile exhibited significantly higher CFI (0.81 ± 0.06) compared to the LA profile (0.69 ± 0.06; $p < 0.001$).
• Component Metrics: MA showed:
  • Higher Relative Amplitude (RA) ($p < 0.001$).
  • Lower Intradaily Variability (IV) ($p < 0.001$), indicating less fragmentation.
  • Higher Interdaily Stability (IS) ($p < 0.001$), indicating more regularity.
  • Higher Cosinor Amplitude and Mesor.

Circadian Phase and Sleep

• Phase: The MA group had an earlier L5c (04:31 vs. 04:59; $p_{adj} = 0.04$), suggesting an earlier circadian phase.
• Sleep Quality: Surprisingly, there were no significant differences in sleep duration, efficiency, WASO, or SRI between the MA and LA groups. This suggests that while movement profiles strongly influence circadian rhythmicity, they may not directly dictate sleep quality metrics in this healthy young population.

Temporal Associations

• L5c (Phase): Inversely associated with early-morning physical activity (06:00–09:00) and late-morning light exposure (09:00–12:00).
• CFI (Integrity): Positively associated with early-evening physical activity (18:00–21:00). It was negatively associated with belonging to the LA profile.
• Light Exposure: CFI was not significantly associated with light exposure at any time of day, suggesting a dissociation between light-driven phase shifts and activity-driven rhythm robustness in this context.

4. Key Contributions

1. Unsupervised Profiling: Demonstrated the utility of k-means clustering on multidimensional actigraphy data to identify distinct movement phenotypes (MA vs. LA) that are independent of demographic variables.
2. CFI as a Clinical Proxy: Validated the Circadian Function Index (CFI) as a robust, single-value metric for circadian health that effectively differentiates between active and sedentary behavioral profiles in real-world settings.
3. Decoupling Sleep and Rhythms: Provided evidence that sleep quality (homeostasis) and circadian rhythm integrity (timing/robustness) can be disentangled; active individuals had better rhythms but not necessarily "better" sleep metrics (e.g., efficiency) than less active peers.
4. Timing Specificity: Highlighted that evening physical activity is specifically linked to higher circadian integrity (CFI), while morning activity/light is linked to phase advancement (earlier L5c).

5. Significance and Implications

• Clinical Application: The study supports the use of actigraphy-derived movement profiles and the CFI as non-invasive, cost-effective screening tools for circadian health. This is particularly relevant for populations where melatonin testing (DLMO) is impractical.
• Therapeutic Strategy: The findings reinforce physical activity as a potent non-pharmacological zeitgeber. Specifically, the association between evening activity and higher CFI suggests that structured exercise in the late afternoon/evening may be a viable strategy to strengthen circadian rhythms, potentially aiding in the treatment of circadian disorders or metabolic issues.
• Public Health: In urban environments where light exposure is often suboptimal, promoting diverse movement profiles (specifically increasing MVPA and reducing fragmentation) could serve as a primary intervention to improve overall circadian health and reduce risks associated with circadian disruption (e.g., metabolic syndrome, cardiovascular disease).
• Future Directions: The authors call for longitudinal studies to establish causality and research into diverse age groups and clinical populations to determine if these movement-circadian relationships hold across the lifespan.