A two-step inverse solution for a single dipole cardiac source

Area of Science:

• Biomedical Engineering
• Cardiovascular Electrophysiology
• Computational Medicine

Background:

• The inverse problem of electrocardiography aims to localize cardiac electrical sources noninvasively using torso potential recordings.
• Optimal electrode configuration for accurate inverse solutions, particularly for premature ventricular contractions (PVCs), remains an open challenge.

Purpose of the Study:

• To investigate the significance of individual torso electrodes for patient-specific inverse solutions of cardiac sources.
• To evaluate the impact of using a reduced, highly significant subset of electrodes on the accuracy of PVC origin localization.

Main Methods:

• A two-step inverse solution was applied to body surface potential data from 128 electrodes in 13 patients with PVCs.
• A greedy algorithm identified electrode significance based on singular values of the transfer matrix.
• Localization error (LE) was calculated comparing estimated and ground truth PVC origins using reduced (32, 64) and full (128) electrode sets.

Main Results:

• The average LE using all 128 electrodes was 28.8 ± 11.9 mm.
• Reduced electrode sets (32 or 64) showed comparable average LEs across different significance criteria, ranging from 28.8 mm to 32.6 mm.
• Inter-patient variability in electrode significance was observed.

Conclusions:

• Accurate localization of a single dipole cardiac source, like a PVC, is achievable with a carefully selected subset of electrodes.
• Patient-specific electrode significance analysis can optimize the noninvasive localization of cardiac electrical activity.