The science

Built on peer-reviewed research,
not wellness folklore

Every metric Trana computes is derived from published studies in top-tier medical journals. Here's the science behind the score.

The cosinor model

At the heart of Trana is the cosinor model — a mathematical technique that fits a cosine curve to your 24-hour heart rate and activity data. This curve captures three key properties of your rhythm:

Amplitude

How tall the wave is — the strength of your rhythm

Acrophase

When the peak occurs — the timing of your biological day

MESOR

The midline — your rhythm's average level

By fitting this model to each day's data, Trana can track how your rhythm shifts over time — and alert you when those shifts cross thresholds that research links to health risks.

Five circadian metrics

Tap each metric to understand what it measures and why.

Relative Amplitude (RA)

Range: 0 – 1

Measures the difference between your most active 10-hour period (M10) and your least active 5-hour period (L5). Higher RA means a strong, well-defined rhythm — your body clearly distinguishes day from night.

Why it matters: Low RA is associated with metabolic syndrome, depression, and cognitive decline. It's one of the strongest non-parametric circadian markers.

Source: Kim et al., JMIR Medical Informatics (2025)

Healthy

0.85 – 1.0

Watch

Below 0.60

Interdaily Stability (IS)

Range: 0 – 1

Measures how similar your rhythm is from one day to the next. High IS means your body clock keeps a consistent schedule — you wake, peak, and wind down at roughly the same times.

Why it matters: Low IS (irregular day-to-day patterns) is linked to cardiovascular disease, hypertension, and obesity in large population studies.

Source: Makarem et al., J Am Heart Assoc (2024)

Healthy

0.60 – 1.0

Watch

Below 0.40

Intradaily Variability (IV)

Range: 0 – 2+

Measures fragmentation within a single day — how often your activity level swings between high and low. Lower IV means a smooth, consolidated rhythm.

Why it matters: High IV (fragmented rhythm) suggests your body clock is struggling to maintain coherent rest/activity cycles. Common in shift workers and those with disrupted sleep.

Source: Makarem et al., J Am Heart Assoc (2024)

Healthy

Below 0.50

Watch

Above 1.0

Acrophase

Range: Time of day

The clock time when your fitted circadian curve reaches its peak. This is your body's biological midday — when alertness, cardiovascular function, and metabolic activity are highest.

Why it matters: Sustained shifts in acrophase (drift) are a leading indicator of mood episodes. Phase delays predict depression; phase advances predict mania.

Source: Lim et al., npj Digital Medicine (2024)

Healthy

Consistent ±15 min

Watch

Shifting 20+ min for 3+ days

MESOR

Range: bpm / counts

The Midline Estimating Statistic of Rhythm — the average value around which your circadian curve oscillates. Think of it as your rhythm's center of gravity.

Why it matters: Changes in MESOR can indicate overall shifts in physiological load, fitness, or recovery state. It contextualizes the amplitude and acrophase.

Source: Shim, Fleisch & Barata, npj Digital Medicine (2024)

Healthy

Stable over weeks

Watch

Sudden sustained change

Source papers

Circadian Complexity Entropy and relative amplitude of heart rate predict metabolic syndrome

Kim et al. · JMIR Medical Informatics · 2025

CCE and RA from wearable heart rate outperform traditional sleep markers for metabolic syndrome prediction.

2025

CosinorAge — biological age from 7-day wearable circadian data

Shim, Fleisch & Barata · npj Digital Medicine · 2024

+1 year of CosinorAge = 8–12% higher all-cause mortality risk in a study of 80,000 participants.

2024

Wearable sleep-wake features predict next-day mood episodes

Lim et al. · npj Digital Medicine · 2024

AUC 0.80–0.98 prediction accuracy. Circadian phase delays predict depression; advances predict mania.

2024

Rest-activity rhythm disruption and cardiovascular disease

Makarem et al. · J Am Heart Assoc · 2024

Rhythm disruption associated with CVD, hypertension, and obesity in a nationally representative US sample.

2024

Irregular circadian patterns and type 2 diabetes

Windred et al. · Lancet Regional Health · 2024

13 million hours of sensor data: irregular circadian patterns associated with T2D incidence.

2024

Circadian disruption and depression in physicians

Real-world MD data · npj Digital Medicine · 2024

50,000+ days of wearable data from 800 physicians: bidirectional link between circadian disruption and depression.

2024

What Trana does not do

Transparency is trust. Here's what we are — and what we are not.

  • Trana is not a medical device and does not diagnose any condition
  • Trana is not a sleep tracker — sleep is one input, not the output
  • Trana does not replace clinical care — alerts are informational, not medical recommendations
  • Trana does not store health data in the cloud — all processing is on-device
  • Trana does not use HealthKit data for advertising or share it with third parties
  • Trana does not claim to predict specific disease — it surfaces patterns that research associates with risk