Video summary

Fitbit Air - One Month Of Scientific Testing!

Main summary

Key takeaways

Science and Nature

Scientific concepts / nature phenomena covered

Biometric heart-rate measurement during exercise

  • Comparison setup
    • A consumer wearable (Fitbit Air) versus a reference ECG chest strap (Polar H10).
  • Primary performance metric
    • Correlation (R) between the wearable and reference measurements.
  • Core idea
    • Placement affects signal quality: the biceps provides cleaner heart-rate signals than the wrist.

Heart-rate monitoring under different exercise modalities

  • Indoor cycling (relatively easy)
    • Fitbit tracks heart rate very closely to the ECG reference.
  • Running (still relatively easy)
    • High correlation overall; biceps slightly improves detection details.
  • Biking/cycling (harder)
    • Wrist signal degrades (attributed to arm tension and handlebar grip).
    • Biceps improves accuracy markedly.
  • Weightlifting (hardest)
    • Wrist often misses high heart-rate peaks.
    • Biceps detects peaks more accurately.

Night-time physiological measurements

  • Heart rate during sleep
    • The section focuses on heart rate (even though the app label mentions HRV).
  • Heart Rate Variability (HRV)
    • Compared using values from the Elite HRV app versus Fitbit estimates.
  • Sleep stages
    • Deep, light, and REM tracked using:
      • Fitbit Air for sleep staging
      • EEG reference: ZMax EG device for scientific sleep-stage tracking

Sleep-stage tracking evaluation

  • Agreement quantified via:
    • Percent agreement/sensitivity
    • An overall “sleep score”
  • Positioning
    • Compared against other brands/models (e.g., Whoop, Apple Watch, Oura Ring, Eight Sleep Pod, and Google/Fitbit/Pixel Watch).

Findings / discoveries reported

Heart-rate tracking (exercise)

Indoor cycling

  • Wrist: correlation ~0.99
  • Biceps: correlation 1.00 (perfect agreement)

Running

  • Wrist: correlation ~0.99
  • Biceps: correlation ~0.99–1.00
    • Improvement described as smaller than in cycling.
  • Wrist occasionally misses portions of heart-rate dips, but differences are characterized as minor.

Biking/cycling

  • Wrist: correlation ~0.92 (reduced accuracy)
  • Biceps: correlation ~0.98
  • Proposed cause: arm tension and wrist movement during gripping the handlebars.

Weightlifting

  • Wrist: correlation ~0.93
  • Biceps: correlation ~0.95
  • Wrist often under-detects high heart-rate peaks; biceps more often captures the full peak.

Hiking

  • Wrist: correlation ~0.88 (weaker / “weirdly” poor initially)
  • Biceps: “almost perfect” (substantially better correlation)

Comparison overview (percentile vs prior tested devices)

Fitbit Air — wrist (example values)

  • Indoor cycling: ~64th percentile, R ~0.99
  • Outdoor cycling: ~57th percentile, R ~0.92
  • Running: ~80th percentile, R ~0.99
  • Weightlifting: ~71st percentile, R ~0.93

Fitbit Air — biceps (example values)

  • Indoor cycling: ~94th percentile, R ~1.00
  • Outdoor cycling: ~83rd percentile, R ~0.98
  • Running: ~94th percentile, R ~1.00
  • Weightlifting: ~83rd percentile, R ~0.97

Overall claim

  • Best performance when worn on the biceps.
  • On the wrist, performance is described as decent but not perfect.

Night-time heart rate and HRV

Heart rate during the night

  • Compared to the reference across roughly ~40–70 bpm.
  • Wrist: correlation R ~0.89 (reported R² ~0.8)
  • Biceps: similar or slightly worse; Fitbit may estimate slightly higher heart rates by a few bpm.

Heart Rate Variability (HRV)

  • Strong correlation for most points on both placements.
  • A small number of outliers; no major wrist vs biceps difference.
  • Conclusion: HRV measurements are considered trustworthy.

Sleep stage tracking (EEG reference via ZMax EG)

Agreement / sensitivity estimates

  • Wrist
    • Deep sleep: ~85%
    • Light sleep: ~80%
    • REM: ~73%
  • Biceps
    • Deep sleep: ~82%
    • Light sleep: ~75%
    • REM: ~76%

Interpretation

  • Both placements perform similarly overall; differences are not described as large.

Overall sleep score

  • Wrist: 79 (“great”)
  • Biceps: 78 (“great”)

Sleep-stage performance relative to other devices

  • Fitbit Air described as among top performers, near:
    • Pixel Watch 4 (new algorithm)
    • Whoop
    • Apple Watch
    • Oura Ring
    • Eight Sleep Pod
    • Google-brand devices grouped as top-tier
  • Fitbit Air is presented as a cheapest way to get near-top sleep-stage tracking.

Methodology / experimental design (as described)

  • Collect one month of data using Fitbit Air worn in two positions:
    • Wrist
    • Biceps
  • Reference instruments used for validation:
    • Polar H10 ECG chest strap for heart-rate accuracy during exercise
    • ZMax EG EEG device for sleep stage validation
  • Performance evaluation metrics:
    • Correlation (R) between reference and Fitbit estimates for heart rate and HRV
    • Percent agreement/sensitivity for sleep stage proportions (deep/light/REM)
    • “Sleep score” computed from average sensitivity across sleep stages
    • Benchmarking against devices previously tested by the reviewer using percentile ranking

Featured researchers / sources (mentioned at end or in subtitles)

  • Rob
    • Channel creator; post-doctoral scientist specializing in biological data analysis
  • Polar H10
    • ECG reference device (research instrument, not a person)
  • ZMax EG
    • EEG reference device for sleep stages
  • Sleep-stage/comparison brands/models mentioned:
    • Whoop
    • Apple Watch
    • Oura Ring
    • Eight Sleep Pod
    • Google devices (including Fitbit and Pixel Watch 4)
    • Amazfit Bip Max
    • Amazfit Helios strap
  • Also mentioned in context:
    • Oura Ring, Whoop Strap, older Fitbit Inspire 3, Fitbit Charge 6

Original video