How to Read a Poincaré Plot

An HRV Poincaré plot places each accepted RR interval on the horizontal axis and the next interval on the vertical axis. The point cloud shows how consecutive beat intervals relate to one another. Its width, length, density and outliers can reveal data patterns, but the plot cannot diagnose a condition by itself.

Start With the Axes

A point at 900 milliseconds on the horizontal axis and 940 milliseconds on the vertical axis means one interval lasted 900 milliseconds and the following interval lasted 940 milliseconds. The next plotted point uses that 940-millisecond interval on the horizontal axis and the subsequent interval on the vertical axis.

The Line of Identity

The diagonal line from lower left to upper right is the line of identity, where two consecutive intervals are equal. Points above it represent a following interval that is longer than the preceding one. Points below it represent a following interval that is shorter.

What SD1 and SD2 Describe

SD1 describes dispersion perpendicular to the identity line and represents short-term point-to-point variation. For a standard lag-one plot using the same accepted intervals and conventional sample calculations, SD1 is mathematically related to RMSSD. SD2 describes dispersion along the identity line and reflects longer variation within the analyzed recording. Both values depend on the recording window and preprocessing.

• SD1 is the standard deviation of the point cloud across its short axis
• SD2 is the standard deviation along its long axis
• SD1 divided by SD2 describes the cloud shape but is not a stand-alone balance score
• Ellipse area is a geometric summary and should not replace inspection of the underlying points

Read the Shape Carefully

A compact cloud means consecutive intervals changed less within that sample. A wider cloud means greater short-term variation. Slow, regular oscillation can stretch the cloud, while clusters, branches or isolated points may reflect changing physiology, nonstationary data, artifacts or rhythm irregularity. Similar-looking plots can have different causes, so shape labels should remain descriptive rather than diagnostic.

Outliers and Artifacts

An isolated point may be produced by movement, weak electrode contact, a missed beat, an extra detection or a real irregular interval. A single bad interval often creates two unusual points because it is paired once with the interval before it and once with the interval after it. Compare the plot with the RR time series and the device signal before excluding anything.

A Simple Reading Sequence

1. Confirm the axes, units, recording length and number of accepted intervals.
2. Check whether the main cloud follows the identity line and whether the recording appears stable over time.
3. Review points far from the main cloud against the RR sequence and signal-quality information.
4. Compare SD1 and SD2 only with sessions collected and processed in the same way.
5. Describe the visible pattern and context without assigning a diagnosis.

Important Limitations

Frequently Asked Questions