A breath morphology workbench ResMed EDF

Okay, but why is
monke awake?

Find sustained shifts into awake-like breathing, then look backward for circuit-visible suspects. “Sleep-wake junk” becomes the signal, not the trash.

01 sessions < 30 min discarded02 sessions combined by sleep night03 awakenings stay hypotheses04 unseen is a real answer

DROPEDFHERE

Nothing loaded

Bring the phat dataset.

Drop a DATALOG folder or select individual BRP, PLD, and EVE files. Sessions shorter than 30 minutes disappear first; surviving sessions are combined into noon-to-noon sleep nights.

Current working model / v0.1

Explainable by construction.

Establish local normal

Breath timing, amplitude, duty cycle, and inspiratory shape are compared against the immediately preceding stable breaths.

Find the directional shift

A candidate needs a sustained rise in irregularity—the direction expected for sleep-to-wake, not just any morphology change.

Estimate episode span

After each transition, the detector follows morphology until rhythm and variability return toward the local sleeping baseline.

Walk backward

Leak, flow limitation, central waxing/waning, event flags, long pauses, and pressure movement compete as precursor evidence.