YASAFlaskified v0.8.11 — full PSG scoring platform (respiratory + staging) built on YASA 0.7

[bartromb]

Hi everyone,

Sharing an update on YASAFlaskified, a web-based PSG analysis platform
built on top of YASA's sleep staging.

What's new in v0.8.11:

• Full AASM 2.6 respiratory scoring now working end-to-end: apnoeas,
  hypopnoeas (Rule 1A/1B), arousals, PLM, Cheyne-Stokes
• Running on YASA 0.7 (latest) — staging + respiratory on a full
  overnight EDF in 5–10 minutes
• Five systematic over-counting corrections (post-apnoea baseline,
  SpO₂ cross-contamination, CSR flagging, confidence stratification,
  artefact-flank masking)
• Hilbert phase-angle effort classification, K-complex exclusion,
  CVR arousal confidence boost

psgscoring — the respiratory algorithms are now also available as
a standalone pip-installable library (pure NumPy/SciPy, no GPU):
https://github.com/bartromb/psgscoring

Live instance: www.slaapkliniek.be — upload anonymized EDFs directly.

GitHub: https://github.com/bartromb/YASAFlaskified

A pilot validation study (n=50, vs. consensus RPSGT scoring) is in
preparation. Happy to hear from anyone working on similar pipelines.

[raphaelvallat]

Thank you for the heads up @bartromb ! Fantastic to hear that YASAFlaskified is actively maintained, and congrats on the psgscoring library, which seems very useful. I'm not able to access the http://www.slaapkliniek.be/ link that you provided ("This site can’t be reached").

Thanks again!

[bartromb]

https://slaapkliniek.be or https://www.slaapkliniek.be

[bartromb]

Only https://slaapkliniek.be

[remrama]

@bartromb is it this? https://sleepai.eu/ 👀

[bartromb]

@remrama yes, that's the one! sleepai.eu, sleepai.be and slaapkliniek.be all point to the same instance now — I actually just fixed the redirects 😉

You'll land on a login page — happy to create an account for you if you want to try it out, just drop me an email.

A bit of context: I'm a pulmonologist/sleep physician and I've been running this in my sleep lab for about two weeks now alongside our Polysomnographic Technologists . Early results are encouraging — reasonable correlation on AHI, and the mild/moderate/severe OSAS separation seems to hold up clinically. Nothing validated yet of course, but it's promising enough...

Cheers from Belgium 🇧🇪,
Bart