Adaptive Personal Intelligence Systems
NerMI turns photoplethysmography and facial tracking into a real-time autonomic interface—tracking valence, motivation, and activation per heartbeat, on-device, without black-box models or cloud dependence.
Per heartbeat
Continuous emotional dynamics
PNS · α-SNS · β-SNS
Three-dimensional emotion
Standard PPG
Research & consumer devices
Personality over time
Lawful physiology
On-device
C-optimized · Cloud-optional
NerMI decomposes the PPG signal into lawful physiological components, mapping directly onto parasympathetic, α-sympathetic, and β-sympathetic activity—no guessing, no generic "stress scores."
Single PPG input → multi-dimensional output
Tracks comfort vs. discomfort through vagal-linked patterns in low-frequency PPG dynamics, acting as a continuous "pleasantness" signal rather than post-hoc sentiment.
Aligned to EMG (zygomaticus–corrugator)
Seconds → weeks
Captures approach-vs-avoidance by extracting mid-frequency oscillations tied to effort, readiness, and behavioral drive—what moves a person toward or away from goals.
Persistent baseline activation
Minutes → weeks
Measures autonomic arousal from heart-rate variability in the ultra-high-frequency band, marking the body's readiness for action—whether excited or stressed.
Short-term autonomic bursts
Seconds → hours
An extended Kalman filter tracks band-limited oscillations in vascular and cardiac changes are integrated with continuous heart rate variability, enabling emotion, state, and trait inference using the same model.
On-device · C-optimized · Cloud-optional
NerMI converts standard PPG into a continuous map of autonomic behavior.
Emotion: heartbeat-level updates in Valence, Motivation, and Activation for moment-to-moment changes.
States: work sessions, interactions, recovery windows, and sleep stages become structured patterns.
Traits: repeated behavioral signatures stabilize into temperament and personality profiles.
Instead of throwing data at a deep net, NerMI encodes the physiology directly. That keeps the logic visible and the computation light.
Each dimension maps to a known autonomic mechanism, not a hidden layer.
EKF frequency decomposition, heart rate variability, and regression models can be inspected, audited, and replicated.
If a device exposes PPG with IBI and pulse volume, NerMI can run on it.
Designed for constrained wearables, not server farms.
Long-term emotional signatures roll up into personality estimates, without questionnaires.
Detect overload, chronic strain, and recovery quality from the same stream used for heart rate.
Interfaces that respond to how a person is actually doing, not just what they click.
Bring quantitative emotion and personality into lab, clinic, and field studies with one pipeline.
