Convergence analysis of the harmonic $B_z$ algorithm with single injection current in MREIT

TL;DR

This paper rigorously analyzes the convergence of the single current harmonic Bz algorithm in MREIT, demonstrating it converges to the true conductivity under mild conditions, supported by numerical validation.

Contribution

It provides the first rigorous mathematical proof of convergence for the single current harmonic Bz algorithm in MREIT, enhancing its theoretical foundation.

Findings

The iterative sequence converges to the true conductivity within an explicit error bound.

Mild conditions on the exact conductivity ensure convergence.

Numerical experiments validate the theoretical convergence results.

Abstract

Magnetic resonance electrical impedance tomography (MREIT) aims to recover the electrical conductivity distribution of an object using partial information of magnetic flux densities inside the tissue which can be measured using an MRI scanner, with the advantage that a higher spatial resolution of conductivity image can be provided than existing EIT techniques involving surface measurements. Traditional MREIT reconstruction algorithms use two data sets obtained with two linearly independent injected currents. However, injection of two currents is often not possible in applications such as transcranial electrical stimulation. Recently, we proposed an iterative conductivity reconstruction algorithm called the single current harmonic $B_z$ algorithm that demonstrated satisfactory performance in numerical and phantom tests. In this paper, we provide a rigorous mathematical analysis of the convergence of the iterative sequence for realizing this algorithm. We prove that, applying some mild conditions on the exact conductivity, the iterative sequence converges to the true solution within an explicit error bound. Such theoretical results substantiate the reasonability and efficiency of the proposed algorithm. We also provide more numerical evidence to validate these theoretical results.