research
Resonaite
A headset reads eight EEG bands once a second. The audio engine folds that reading back into the modulation of the track playing, so the music changes while you are still hearing it. Built primarily by Liqian Yan.
Notice
Research prototype, not a medical device. Resonaite has not undergone clinical validation and is not intended to diagnose, treat, cure, or prevent any disease or condition. Nothing here should be read as a therapeutic claim.
Resonaite plays a track and watches you listen. A consumer EEG headset samples eight frequency bands every second, the engine derives an engagement index and an alpha ratio from them, and those numbers feed straight back into how the music is modulated. The loop closes in real time.
Modulating audio this way usually sounds like a test tone. A seven-stage pipeline fixes that by adding organic jitter, breathing depth, and waveform morphing, so the effect stays musical. The figure below is the real output of that pipeline, not an illustration.
The rest is ordinary engineering done properly: 73 curated tracks normalised to a single loudness standard, new tracks generated on demand through AI music models with tempo and key enforced, and an offline-capable progressive web app so a session survives losing the network.
Created primarily by Liqian (Eric) Yan, Co-Founder & CFO.
Signal pipeline
7-stage CIM engine
EEG
NeuroSky MindWave, 8 bands at 1 Hz
Library
73 tracks at −14 LUFS
Client
Offline-capable PWA
Built with
