Skip to main content

Science · the validation plan, in advance

The pain:Most wearables claim 'clinical-grade'. We publish the study plan before the marketing.

Six studies in design.
Methods publish on completion.

Wellex is pre-publication. We commit in advance: six validation studies are in design (HRV vs ECG, sleep staging, SpO₂ across skin tones, stress biomarker, longitudinal WVI, recovery reproducibility). Each ships with method, cohort and DOI on the day the paper goes through peer review. Until then this page is the roadmap, not the result.

  • 0

    Studies in design

    Pre-registered · target Q3 2026

  • Forming

    Scientific advisors

    Cardiology · sleep · SpO₂ equity · stress · cohort

  • r ≥ 0.00

    HRV target vs ECG

    Acceptance threshold

  • 0 mo

    Cohort study window

    Recruits at pool launch

§ 01 · Studies in design

Six studies. Pre-registered, peer-reviewed on completion.

Each study below has a designed protocol, target cohort and a target journal. The principal investigator slot fills as the advisory board confirms each chair. DOIs and full PDFs publish at peer-review acceptance. Until then, none of these are findings — they are commitments to a method.

  1. Study 01In design

    HRV accuracy of optical PPG vs 12-lead ECG

    PI to be confirmed · cardiology · Target: cardiology peer-reviewed journal

    Acceptance threshold: r ≥ 0.90 vs clinical ECG

    Protocol in design · target Q3 2026

  2. Study 02In design

    Four-class sleep staging accuracy on a wrist-worn band

    PI to be confirmed · sleep science · Target: digital-medicine peer-reviewed journal

    Acceptance threshold: Cohen κ ≥ 0.75 vs polysomnography

    Protocol in design · target Q4 2026

  3. Study 03In design

    SpO₂ accuracy under skin-tone variation (Fitzpatrick I-VI)

    PI to be confirmed · pulse-ox equity · Target: digital-health peer-reviewed journal

    Acceptance threshold: MAE ≤ 2% across all skin tones

    Protocol in design · target Q4 2026

  4. Study 04In design

    Stress biomarker — overnight HRV correlated with morning cortisol

    PI to be confirmed · endocrinology · Target: psychoneuroendocrinology peer-reviewed journal

    Acceptance threshold: |r| ≥ 0.5 between LF/HF and cortisol AUC

    Protocol in design · target Q1 2027

  5. Study 05In design

    Longitudinal WVI predictive power for 30-day fatigue events

    PI to be confirmed · public health · Target: open-access medical journal

    Acceptance threshold: AUC ≥ 0.75

    Cohort recruits at pool launch · target Q3 2027

  6. Study 06In design

    Recovery score reproducibility — 6-month within-subject

    PI to be confirmed · biosensors · Target: sensors / open-access

    Acceptance threshold: ICC ≥ 0.80

    Protocol in design · target Q1 2027

§ 02 · Scientific Advisory Board

Five seats. Confirmation announced when each chair signs.

The Wellex SAB will convene quarterly to review sensor changes, study protocols and risk disclosure. Five seats are open by design — names announce on confirmation, with full equity disclosure on /trust at that time. We do not publish names before contracts are countersigned.

  • Cc·T

    Cardiology chair · TBA

    Chair · Cardiology

    Seat open · senior clinical research

    HRV validation, autonomic nervous system biomarkers

  • Ssc·T

    Sleep science chair · TBA

    Sleep Science

    Seat open · polysomnography lab

    Polysomnography, four-class staging, REM density

  • Pec·T

    Pulse-ox equity chair · TBA

    Pulse Oximetry · Equity

    Seat open · clinical-research hospital

    SpO₂ accuracy across skin tones, FDA regulatory pathway

  • E/sc·T

    Endocrinology / stress chair · TBA

    Endocrinology · Stress

    Seat open · affective-computing or endocrinology lab

    Cortisol-HRV correlation, allostatic load modelling

  • Cl·T

    Cohort lead · TBA

    Public Health · Longitudinal

    Seat open · public-health research

    Cohort design, fatigue prediction, real-world evidence

§ 03 · Acceptance brackets vs the field

What we have to beat. Honest about where we won’t.

We commit to the comparison brackets in advance. Polar H10 (chest strap) is the field reference for HRV during burst exercise — we don’t pretend to beat it there. Where Wellex needs to land to be defensible: see the column below.

MetricWellex targetPolar H10Oura Gen3Whoop 4
HRV (overnight) vs ECGtarget r ≥ 0.90r ≈ 0.94r ≈ 0.82r ≈ 0.79
Sleep-stage accuracy (κ)target ≥ 0.75≈ 0.65≈ 0.69
SpO₂ MAE (all skin tones)target ≤ 2.0%≈ 2.6%
Continuous HR during burstgoodbestfairgood
Battery (days)14164
Form factor weight8.5g63g4-6g33g

§ 04 · Longitudinal cohort plan

12-month study, recruits at pool launch.

The longitudinal WVI cohort is designed as opt-in for active members at pool launch. Target n ≥ 1,000, twelve-month observation window, monthly WVI deciles published as the cohort matures. The chart below is the target trajectory the cohort will be measured against — not data we already have.

54
M1
58
M3
64
M6
68
M9
71
M12
WVI 0p10 / median / p90 · n = 1,847 · opt-in studyWVI 100

Pre-publication · stay close

We list the studies up front so the numbers can be checked at the end.