Hypertension Coupling Observatory

Hypertension is more than a blood-pressure number.

This is an interactive explainer of cardiovascular computer-model experiments, showing how different simulated causes of high blood pressure change blood flow and pressure waves in the arteries. The controls let you explore saved results; they do not rerun the models, and the charts are not patient measurements or camera signals.

Model-derived demonstration · no participant data

Explore the model

Creation route

Reference site

Inlet baseline

Central stiffness multiplier 1.75×

Mean arterial pressure (MAP)—model operating point

Waveform-shape difference—normalised RMSE

Screen pass rate—illustrative stiffness screen

Flow note—route relationship

Interpretation guardrail
These traces are carotid reference-site pressure waveforms. They are not simulated rPPG signals and do not establish camera performance.

How to explore: Choose a model route, carotid reference site, and inlet waveform, then adjust the stiffness setting. The route choice changes the route readouts; the site, inlet, and stiffness choices change the waveform summaries. These controls do not rerun the models.

Renal flow: route divergence

The dashed line marks the reference value of 1. A bar above it means the model produced more kidney blood flow than the reference; a bar below means less. These ratios come from model outputs, not clinical measurements, and the two route cases are not an exact pressure match.

Reference-site topology

central conduitreference branch

Carotid shape response to central stiffness

Waveform-shape difference (normalised RMSE) measures how different two waveform shapes are: larger values mean more difference. The stiffness settings run from left to right. These are model waveforms, not measurements from a person.

Feature preservation by low-pass cutoff

A low-pass filter removes faster signal changes above its cutoff. Lower cutoffs remove more detail; higher cutoffs retain more. The line shows the screen pass rate across the cutoff settings; it does not describe transfer from artery to skin or camera.

What the model makes visible

01 · ENVELOPE

Starting pressures the model accepts

The standard model may reject some adult blood-pressure values as outside its allowed setup range.

02 · ROUTES

Two ways to set up hypertension

Changing pressure directly and changing cardiovascular properties can produce different simulated flow patterns. In separate pressure-matched model comparisons, the routes still differed in cardiac output, resistance and regional flow.

03 · STABILITY

Model stability can be uneven

Some settings fail to settle while more extreme settings converge. Stability does not change in a simple, predictable direction.

04 · REFLEX

The reflex may adapt to new pressure

With default settings, the model can adjust to a pressure change as if that pressure were the new normal.

05 · INLET

The inlet waveform has a limitation

Pulse’s stock aortic inflow lacks the brief backward-flow phase after ejection that appears in the distributed reference waveform.

06 · P1

Kidney flow shifts with pressure

The resistance adjustment changes kidney flow, but does not keep its share of total flow constant as pressure changes.

07 · P3

Parameter names need checking

In the tested build, the field called “gain” changed how quickly the reflex adjusted, not its final sensitivity.

08 · WAVEFORM

Some stiffness changes alter waveform shape

In these model runs, compliance changes altered the simulated internal-carotid shape above the chosen noise screen. This does not show that a camera can detect it.