Decode the signal beneath thought.
AXON fuses EEG, galvanic skin response, and eye-tracking into a single live cognitive signal — built for researchers mapping attention, stress, and emotion in real environments.
Raw cortical signal, captured at the source.
Dry-electrode EEG channels sample cortical activity at 512 Hz, isolating alpha, beta, and theta bands before a single frame of noise reaches the model.
Three biometric streams, one timeline.
Galvanic skin response and pupil dilation are time-aligned against the EEG stream, giving every cognitive event three independent points of confirmation.
Emotion and attention, decoded live.
A lightweight on-device model classifies engagement, cognitive load, and affective state in under 40 milliseconds — fast enough to close the loop in real time.
From signal to spatial understanding.
Decoded states are projected onto the physical environment — a shelf, a screen, a room — producing a live attention and stress heatmap researchers can act on.
Three sensors. One cognitive signal.
Every module streams into the same synchronized timeline, so researchers see attention, arousal, and gaze as one continuous picture — not three disconnected logs.
EEG tracking, 8–32 channels
Dry-electrode caps stream raw cortical band power in real time, with automated artifact rejection so movement and blink noise never reach the model.
GSR sensors
Wrist-worn electrodermal sensors track sympathetic arousal at 32 Hz, flagging stress spikes the moment they begin.
Eye-tracking heatmaps
Gaze vectors are resolved into live attention heatmaps over any physical or digital surface, down to 0.5° of visual angle.
On-device decoding
Classification runs locally on embedded hardware — no cloud round-trip, no raw biometric data leaving the room.
Environment-aware mapping
Decoded states are projected onto shelving, displays, or workstations, giving retail and service researchers a live map of where attention and stress actually occur.
Privacy-first by architecture
Every session is pseudonymized at capture and aligned to GDPR and EU AI Act requirements for biometric research — consent, retention, and deletion are built into the pipeline, not bolted on.
Signal in. Response out.
Signal Acquisition
EEG, GSR, and eye-tracking hardware begins streaming synchronized raw biometric data the moment a session starts.
AI Emotion Decoding
The fusion model classifies cognitive load, stress, and affective valence from the combined signal in under 40 ms.
Real-Time Heatmap Generation
Decoded attention and stress states are rendered as a live spatial heatmap over the physical or digital environment.
Adaptive Environment Response
Downstream systems — lighting, audio, layout — receive the live signal and adjust in real time to the observed state.