Skip to content

docs: add PSG EEG validation path for sleep apnea research #879

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
2023Anita wants to merge 1 commit into
base: main
Choose a base branch
from
Open

docs: add PSG EEG validation path for sleep apnea research #879

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
16 changes: 16 additions & 0 deletions docs/edge-modules/medical.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -145,6 +145,22 @@ println!("AHI: {:.1}", detector.ahi());

6. **Limitations**: WiFi-based breathing detection works best when the subject is relatively still (sleeping). Tossing and turning may cause momentary breathing detection loss, which could either mask or falsely trigger apnea events. A single-night study should always be confirmed with clinical polysomnography.

#### Research Extension: PSG/EEG Co-Validation

For sleep-medicine research, the sleep-apnea module should be validated against a reference sleep study rather than interpreted in isolation. A practical co-validation design is:

1. **Reference channel**: collect polysomnography (PSG) or an approved sleep-study export with EEG, EOG, EMG, airflow, respiratory effort, pulse oximetry, body position, and scored sleep stages where available.
2. **RuView channel**: record time-synchronized CSI-derived breathing rate, motion energy, presence confidence, apnea start/end events, and AHI updates from the same overnight window.
3. **Clock alignment**: align both systems with a shared start marker, wall-clock timestamp, or manually annotated synchronization event before comparing epochs.
4. **Epoch-level comparison**: compare RuView apnea candidates against scored respiratory events in 30-second windows, and stratify false positives by sleep stage, body position, and motion artifact.
5. **Clinical review loop**: have a sleep-trained clinician review discordant epochs before changing thresholds, because EEG arousals, wake-after-sleep-onset, REM/NREM stage, and movement artifact can explain apparent disagreement.

This approach keeps RuView in its intended role: an experimental, contactless sensing layer that can be studied against PSG/EEG-derived ground truth. It should not replace PSG, HSAT, clinician scoring, or regulated monitoring equipment.

Related references:
- AASM Scoring Manual overview: https://learn.aasm.org/AssetListing/The-AASM-Manual-for-the-Scoring-of-Sleep-and-Associated-Events-4265/The-AASM-Manual-for-the-Scoring-of-Sleep-and-Associated-Events-6697
- AASM Practice Standards: https://aasm.org/clinical-resources/practice-standards/

---

### Cardiac Arrhythmia Detection (`med_cardiac_arrhythmia.rs`)
Expand Down