A Bi-objective Array Optimization Framework for Magnetocardiographic Source Imaging.

Abstract

The rapid development of weak magnetic sensing technologies, most notably optically pumped magnetometers (OPMs), has enabled flexible and reconfigurable magnetocardiographic source imaging (MCSI) systems. As a paradigmatic distributed sensor platform, MCSI relies critically on the spatial arrangement of sensors, which decisively determines both system-level cost and achievable source reconstruction performance. Nevertheless, principled array design strategies for MCSI remain an unmet need. This study aims to propose a novel strategy for sensor array design in MCSI systems. This study proposes, for the first time to our knowledge, a bi-objective array optimization framework (BoAOF). The framework models array design as a dual-objective optimization problem aimed at maximizing array sensitivity while minimizing the total spatial deviation. It integrates forward modeling with the non-dominated sorting genetic algorithm II (NSGA-II) to generate a Pareto front of sensor configurations, followed by a hybrid CRITIC-Entropy-VIKOR decision process to identify the optimal compromise solution. In both simulation and phantom experiments, the arrays optimized via the BoAOF consistently outperformed competing baseline methods, achieving lower mean dipole localization error, reduced mean spatial deviation, and higher mean source‑reconstruction signal‑to‑noise ratio across a range of channel counts and source strengths. This study proposes a novel strategy for sensor array design in MCSI systems. This work provides a methodological basis for the design of flexible OPM-based MCSI systems.