ALIGNED // LIVE148 BPM
Cardiac
Heart rate, rhythm and variability—read as a pattern over time, not a single number.
InSIDE™Request a briefing 01 The platform
A patient is a timeline,
not a collection of tiles.
InSIDE™ is being built to bring the signals already present around a neonatal bed into one coherent, time-aware view. The goal is simple: help a care team see a change taking shape, not just the alarm it eventually triggers.
Discuss a deployment ↗Research-stage systemMultimodal · Edge-first · Human-led
InSIDE™ is designed around the equipment, interruptions and clinical judgment already present at the bedside—not an idealized lab.
U.S. Navy / MC3 Jake Berenguer · Public domain ↗02 One evolving picture
Signals become useful
when they share context.
A brief oxygen dip means something different when a baby has just turned over. A temperature change means something different when perfusion and heart-rate behavior are shifting at the same time.
InSIDE™ preserves those relationships. Each stream is time-aligned, checked for quality and attached to a longitudinal patient model. The system is meant to support a clinician’s question—what changed, and what changed with it?—without pretending there is one magic number.
03 The signal fabric
ALIGNED // LIVE96 % SpO₂
Oxygenation
SpO₂ trends are interpreted alongside motion, respiratory behavior and signal quality.
ALIGNED // LIVE36.8 °C
Thermal
Non-contact temperature maps can surface peripheral change and evolving spatial patterns.
ALIGNED // LIVELOW STATE
Movement
Position, activity and stillness help separate physiological change from ordinary motion artifact.
ALIGNED // LIVE30 FPS
Video
Visual features add respiratory and behavioral context without becoming the clinical record.
ALIGNED // LIVE06 EVENTS
Clinical events
Interventions and observations give the model the context required to make a trend meaningful.
04 From bedside to decision
Four moves.
No black box.
Every stage has a job, an owner and a boundary. That is how a research model becomes something a hospital can evaluate responsibly.
- 01LIVE
Capture
Collect continuous physiological, thermal and behavioral streams using modular bedside hardware.
- 02EDGE
Align
Synchronize timestamps, flag signal quality and separate identity before analytical processing.
- 03TIME SERIES
Model
Compare current trajectories with the patient’s own baseline and clinically labelled outcomes.
- 04HUMAN REVIEW
Explain
Show the signals and events behind a risk change so the care team can investigate in context.
05 Designed for the room
A live, readable trajectory.
Care teams can move from a warning to the exact sequence of physiological and observed events that preceded it.
A common operating picture.
At-a-glance views help teams see which patients are stable, changing or need a closer look—without replacing bedside assessment.
An evidence trail.
Predictions, labels and outcomes can be reviewed together, making model performance auditable during validation.
Built for clinical validation