SpiroSmart

[sportsrancho]

homes.cs.washington.edu/~mayank/Papers/TrackingLungFunction.pdf

[mnholdem]

Nice detective work!

Excerpt:

[2]. SpiroSmart [1] is a smartphone-based approach that measures
many common lung function measures using the phone’s built-in
microphone. It does not require any additional hardware or
adapter to be attached to the phone. The low cost of smartphones
as compared to spirometers allows lowering of access barriers to
full-fledged lung function measurement in the developing world.
SpiroSmart has a median error of 8.01% for the most common
lung function measures and has been proven to the effective for
diagnosing presence of lung ailments. However, the original
design required a smartphone. In order to bring this technology
within reach of a global population, we are working on a call-in
service based system (coined SpiroCall), which is agnostic to the
phone type. Users will be able to simply call in to a server with
their personal phone and record their exhalation data. Our
preliminary results show that the loss of bandwidth and resolution
due to data transmission through voice channel does not have
significant adverse effect on the performance, vastly reducing the
access barriers for a technology such as SpiroSmart.

2. TECHNOLOGY
SpiroSmart requires users to hold the phone at approximate arm’s
length, inhale their full lung capacity, and then forcefully exhale
at the screen of the phone until the entire lung volume is expelled.
The microphone records the exhalation and sends the audio data
to a server, which calculates the flow rate by estimating
physiological model of user’s vocal tract and the reverberation of
sound around the user’s head. The microphone acts as an
uncalibrated pressure sensor. SpiroSmart converts this pressure
into a flow rate by compensating for pressure losses between
phone and mouth, and keying in on acoustical properties of the
lips that relate pressure to flow. Then, a number of features are
extracted from this signal through: (1) envelope detection, (2)
spectrogram processing, and (3) linear predictive coding (LPC).
These features give an estimate of the uncalibrated flow rate. We
then use regression to obtain specific lung function measures from
these approximations. Lung function typically consists of lung
function measures like PEF, FEV1, FEV1% and FVC.

---

This may finally be the proof which I have long suspected: A woman's lung function should be easy to calculate by how often and long she talks on the phone.

---

Seriously, though, this is some very cool science.

[lakon]

Not to sound like a conspiracy nut, but it is out of the University of Washington Seattle, WA. At least, RLS could go check it out easily enough if they are concerned about the nutty FDA putting a spirometer in their way too.

...and is it just me or does the obstructive flow curve look a lot like the Afrezza PK profile?

I'm just saying...