Two layers run in parallel: (1) continuous raw streaming for immediate “voice” motion, and (2) a slower feature layer that shapes larger-scale sonic / spatial behavior.

Data source → streamer → OSC → audio / spatial engine → multichannel output
Polysomnographic recordings (EEG + EOG/ECG/EMG) are streamed from CSV at the original sampling rate (512 Hz) and routed as individual control streams. Each stream directly modulates a corresponding tone (or parameter cluster), producing a temporally precise multichannel texture.

Raw layer: “each sensor as a voice” (continuous, sample-accurate feel).
Feature layer: band-power contours and event-like triggers (slower updates) that steer timbre, diffusion, and stability.

The interface is designed for exploratory listening—like moving a microscope around an inner landscape.

• Solo / mute / rebalance channels to compare regions (e.g., frontal vs occipital; left vs right).
• Contrast brain vs physiology by emphasizing EEG or isolating EOG / ECG / EMG.
• Stage comparison (wake → N1 → N2 → N3 → REM) by switching excerpts or navigating within a recording.
• Focus listening modes (e.g., “micro-detail” vs “sleep weather”) by blending raw vs feature layers.

To avoid arbitrary one-to-one mappings, the system groups derived features into a small set of perceptual “macro controls” that behave like weather patterns: density, agitation, brightness, stability, depth, and spatial diffusion.

Features are computed on short windows (seconds) with overlapping hops, then streamed via OSC at a slower rate than the raw layer.

This provides coherent, interpretable motion while preserving the raw microstructure in parallel.

(Mapping is intentionally adjustable: the same features can drive sound, light, spatialization, or haptics, depending on the installation context.)

Auxiliary physiology is treated as an equal musical / physical partner rather than “background data.”

• ECG: can drive beat-triggered events (true heartbeat timing rather than an imposed tempo).
• Inter-beat interval (IBI): can steer spatial diffusion / “breath” or other slow parameters for stable, legible motion.
• EOG: eye movements can be rendered as lateral sweeps or gestural motion (especially resonant in REM).
• EMG: muscle tone can contribute to noise components, texture, or haptic intensity.

Where needed, simple cleanup can be applied for reliability (e.g., peak-debounce for ECG triggers), while keeping the raw streams available for inspection.

This is not a classifier—think of it as a set of listening cues. Sleep is variable; the goal is to make tendencies perceptible.

• Wake: often more alpha/beta presence; faster “edge” motion.
• N1 (drowsy onset): alpha tends to soften; theta often emerges as drifting depth.
• N2: spindles can appear as brief shimmering bursts; K-complexes can read as sudden large-scale gestures.
• N3 / SWS: delta dominance—slow, heavy waves; strong synchronizations can feel like collective breathing.
• REM: mixed-frequency textures; EOG gestures may become more prominent.

The sonification strategy has evolved over the course of the project.

Initial conception uses more downsampled / melodic mappings—musically coherent, but less faithful to fast temporal structure.

Later conception introduces continuous frequency modulation: each channel directly shapes the motion of its own tone in real time. The result is a richly textured, temporally precise choir in which rhythms and synchronizations become audibly present. One aim of this conception is interpretability through embodiment: preserving micro-timing while offering macro “weather” cues that help listeners orient themselves.