# Impact of rigid cardiac motion on the accuracy of electrocardiographic imaging

## TL;DR

This paper investigates how heart movement affects the accuracy of non-invasive cardiac electrical activity reconstruction using ECGI.

## Contribution

The study introduces a systematic analysis of how different types of rigid cardiac motion affect ECGI accuracy.

## Key findings

- Rotational heart motion has a stronger impact on ECGI accuracy than translational motion.
- Pitch rotation causes the largest changes in ECGI accuracy, while y-axis translation has the least impact.
- ECGI is robust to small heart displacements, with minimal changes in accuracy for 10 mm or 10° movements.

## Abstract

Electrocardiographic Imaging (ECGI) offers a non-invasive approach to reconstruct cardiac electrical activity. However, the inverse problem of ECGI is highly ill-conditioned, making it sensitive to errors. In practice, rigid displacements of the heart during beating introduce geometric errors into the ECGI problem. This study aims to investigate the impact of cardiac rigid motion on the accuracy of ECGI.

We employed the Boundary Element Method (BEM) to solve the forward problem and the Tikhonov method to address the inverse problem. We utilized a dataset from the CRVTI/SCI Institute, which involves Langendorff-perfused dog hearts suspended in a torso-shaped tank. Based on clinical experience, six different types of cardiac movement patterns, including translations and rotations, were designed to assess the impact of various displacements on the accuracy of the ECGI solution.

Our study found that among the translational and rotational movements of the heart, rotational motion should be prioritized for attention, as it caused significantly stronger changes in ECGI correlation coefficient (CC) and relative error (RE) than translational motion. Among the translations along the coordinate axes, movement along the y-axis (anterior-posterior movement within the chest cavity) had the least impact. For rotational movements, rolling had the least impact, yaw had moderate impact, and pitch had the greatest impact.

The inverse solution of ECGI demonstrates a certain robustness to changes in heart position, with CC changes of less than 2% for 10 mm displacements and less than 5% for 10° rotations. This suggests that ECGI changes due to cardiac geometric motion can be disregarded within a certain range.