Run experiments on the world’s richest ANS dataset
Query thousands of verified HRV recordings, derive novel biomarkers, and contribute models back to the community — all on consent-driven, anonymised data.
Reflex Lab is built on active clinical research collaborations and government-backed innovation funding — combining medical-grade validation with the resources to scale a new class of ANS biomarker.
Clinical Research
NHS Trusts
Active research collaborations validating Reflex HRV signals against clinical gold standards inreal care settings.
Innovation Funding
Innovate UK
UKRI-backed grant funding supporting Reflex's R&D in autonomic nervous system biomarker discovery and wearable validation.
Data Pipeline
From wrist to query
REFLEX wearables capture raw RR intervals at medical grade. The data flows into the Bio Data Bank and is exposed via a governed query layer — with full audit trails and on-chain attribution.
Activity type, device model, firmware, session quality score. All exports are fully anonymised.
Evidence · Method comparison
Measured against clinical references
Before a metric earns a place in Reflex, it's checked against an established reference — Vive ring-derived inter-beat intervals against a Polar H10 chest-strap ECG, the gold standard for R–R timing. Every point below falls inside the ±5 bpm corridor used as the industry tolerance.
Vive ring — measuredPolar H10 ECG — reference & tolerance
Fig. 01 · Heart rate agreementrest · stand · motion
Readoutref: Polar H10
0.91bpmMean absolute errorHR vs ECG
0.97Correlation rlinear agreement
100%Within ±5 bpmtolerance band
4msMedian IBI errorR–R interval
~2msRMSSD agreementshort-term HRV
Heart rate, line of identity. Across rest, standing and motion, ring-derived HR tracks the ECG reference to a mean absolute error of 0.91 bpm (r = 0.97); every sample lands inside the ±5 bpm corridor.
R–R intervals & HRV. The Vive ring reproduces ECG beat-to-beat intervals to a median error of 4 ms (r = 0.97–0.98), with RMSSD agreeing to within ~2 ms — differences sit within the limits of agreement with negligible bias.
Single user · single testReference: Polar H10 ECGCharts illustrate agreement; larger-n results to follow
Access
Free to access. Pay only for what you query
Sign up and start querying immediately — no subscription, no upfront commitment. Access is free for all approved researchers.
You’re only charged for the data you pull. Bulk and cohort queries are supported at any scale.