ECG results as a colour map? A project by WUT Faculty of Physics researchers

A multi-hour ECG examination produces a recording that encompasses thousands of heartbeats. A conventional graph allows individual beats to be analysed in detail, but reviewing the entire recording is far more difficult. Heart-rate variability plots can help by showing the intervals between successive beats. However, they do not reveal whether the shape of the beats changes. The method developed by the researchers combines both types of information in a single image, which they call a carpet plot. The program identifies successive R waves in the ECG—the signal’s characteristic peaks—and selects a section of the recording around each one. It then arranges these sections in successive rows and represents voltage values using colours.

"We look at the electrocardiogram through the eyes of physicists and see a quasiperiodic signal that reflects the dynamics of a complex system: the heart. These dynamics comprise processes that are both local, such as the depolarisation and repolarisation of cardiomyocytes, and global, such as heart-rate variability and coupling with respiration. Using two time scales simultaneously therefore seemed natural to us," says Tomasz Gradowski, PhD.

On a map prepared in this way, successive rows correspond to successive heartbeats. Sections of the recording are neither stretched nor compressed, so the actual RR intervals associated with the instantaneous heart rate are preserved. At the same time, the colours show changes in the morphology of the QRS complex, ST segment and T wave. When the heart beats regularly, ordered bands form on the map. When an arrhythmia or a change in morphology occurs in the recording, this pattern changes.

The method was tested on recordings from several publicly available databases covering healthy individuals and patients with various cardiac disorders. The plots revealed, among other things, episodes of atrial fibrillation, premature ventricular contractions, long pauses in the rhythm and atrioventricular block. In a patient with long QT syndrome, changes in the QT interval and T-wave morphology could be observed alongside changes in heart rate. In an exercise test, distinct patterns were formed by rest, exercise and recovery.

The method is described in the journal ‘Computer Methods and Programs in Biomedicine’ in the article ‘A novel method for analysis of transient morphological changes in quasiperiodic physiological signals and their neurogenic correlates’. It is not yet a diagnostic tool.

The carpet-plot method is not the only algorithm developed by the Human Cardiovascular System Physics Laboratory, whose members include the authors of the paper. The group of physicists has also recently developed a method for reconstructing unipolar potentials that makes it possible to perform a virtual heart biopsy.