[Equipment] Any experiances with Ringconn Gen 2?

[djbclark]

Anyone have a Ringconn Gen2 and care to speak about its efficacy re: sleep apnea?

It'd be esp. good if you have a traditional device to compare its results with.

As you may know they use AI/Deep Learning to do detection with non-continuous monitoring. They have at least 1 paper on this.

The first paper, "A Transformer-Based Deep Learning Model for Sleep Apnea Detection and Application on RingConn Smart Ring," explores a new method for detecting Obstructive Sleep Apnea Hypopnea Syndrome (OSAHS) using the RingConn Smart Ring. OSAHS is a common sleep disorder characterized by severe snoring, repeated breathing pauses during sleep, excessive daytime sleepiness, morning headaches, and memory loss. Severe cases can also lead to nighttime asphyxia and cardiovascular problems.

Traditional OSAHS detection requires an overnight examination with bulky polysomnography (PSG) equipment in hospitals. This paper proposes a deep learning model based on Transformer, utilizing the multimodal physiological signals of the RingConn Smart Ring to detect OSAHS.

The model achieved excellent detection results on public datasets with an F1 score of 76.6. The study also evaluated the performance of signal combinations in detecting oxygen desaturation events (DESAT) with the RingConn Smart Ring. Notably, no current research validates OSAHS monitoring based on a smart ring. Thus, the study used data collected from the RingConn Smart Ring for verification and performed PSG-AHI estimation.

Results showed a high correlation (ρ=0.96) between the AHI calculated by the RingConn Smart Ring and PSG-AHI, demonstrating the high accuracy and convenience of the RingConn Smart Ring in medical health monitoring.

[LJClark]

I started with the Ringconn Gen 1, and then moved to the Gen 2.

Bottom Line Up Front:  I think that Ringconn's claims for the Gen 2 ring's efficacy related to sleep disorders in general, and OSAHS in particular, are rather optimistic -- and perhaps even speculative.  

Their approach to sleep apnea "detection" is for three consecutive nights with the ring's data acquisition rate greatly increased (the administrator of their Facebook page even mentioned SpO2 data at 1 second intervals).  The AI seems to be primarily focused on SpO2 and Pulse Rate, and may also use an accelerometer along with some undisclosed data.

Normally (not the apnea regime) the Ringconn collects (displays?) SpO2 data at 5 or 30 minute intervals.  Even that information is displayed in either time block ranges or as a curve-smoothed time strip.  

What worries me are comments on the Facebook page from users who seem to be accepting the Ringconn information (I don't think that the data is actually available) as sufficient to inform their medical and therapeutic decision making.

Last night I started the 3 night RingConn sleep apnea routine.  I am simultaneously collecting my normal data from the ResMed 11, supplemented by simultaneously recorded pulse and SpO2 from my Nonin 3150BT (4-second sampling rate).  Both the Ringconn and Nonin are collecting from the same finger.  Ringconn's "high correlation" AHI after the first night is:  ResMed11 + Nonin = 3.99; Ringconn = 0.4.  I have two more nights -- I usually take 3 nights any changes or enhanced observations.

BTW -- The Nonin 3150 has been a revelation -- adding two OSCAR data streams which are not dependent on the C/APAP's air flow.

I'm looking seriously at Oura and Ultrahuman rings.  My gut reaction to Ringconn's claims is not at all positive.  But all these companies rely on the potential purchaser being satisfied with marketing-level information, not on the actual sampling rates, data acquisition and handling, etc.  (Some people get upset with Oura's subscription model.  The monthly cost is less than a 1/4-pouder with cheese and a medium drink.)